Iron Age Cremation Burials in South- Eastern and West Estonia

TALLINNA ÜLIKOOL HUMANITAARTEADUSTE DISSERTATSIOONID

TALLINN UNIVERSITY DISSERTATIONS ON HUMANITIES

RAILI ALLMÄE

IRON AGE CREMATION BURIALS IN SOUTH- EASTERN AND WEST ESTONIA. AN OSTEOLOGICAL APPROACH

Tallinn 2017

3 TALLINNA ÜLIKOOL HUMANITAARTEADUSTE DISSERTATSIOONID TALLINN UNIVERSITY DISSERTATIONS ON HUMANITIES

Raili Allmäe IRON AGE CREMATION BURIALS IN SOUTH-EASTERN AND WEST ESTONIA. AN OSTEOLOGICAL APPROACH Institute of Humanities, Tallinn University, Tallinn, Estonia

The dissertation was accepted for the defence of the degree of Doctor Philosphiae in History by the Doctoral Studies Council of Humanities of Tallinn University on April 18th, 2017.

Supervisors: Marika Mägi, Senior Researcher at Tallinn University Leiu Heapost, PhD Opponents: Gunita Zarina, Senior Researcher at Latvian University Laurynas Kurila, Researcher at Lithuanian Institute of History

The defence will take place on June 19th, 2017 at 16 o’clock at Tallinn University lecture hall M-649, Uus-Sadama st 5, Tallinn.

This research was supported by the European Social Fund’s Doctoral Studies and Internationalisation Programme DoRa and by the Estonian Ministry of Education and Research, targeted financing no. SF0042476s03 and SF0130012s08. Also by Estonian Science Foundation grants ETF5973 and ETF6899.

ISSN 1736-5031(pdf) ISBN 978-9949-29-330-8(pdf)

Tallinn University Narva Rd 25 10120 Tallinn www.tlu.ee

4 CONTENTS

LIST OF PUBLICATIONS ...... 7 ABSTRACT ...... 9 PREFACE ...... 10 ACKNOWLEDGEMENTS ...... 11 INTRODUCTION ...... 12 The study of cremated human remains in Northern Europe ...... 13 The osteological research on cremated human remains in Estonia ...... 17 The history of palaeodemographic investigations in Estonia ...... 19 AIMS ...... 22 1. MATERIAL ...... 23 1.1. West Estonia ...... 23 1.2. South-eastern Estonia ...... 25 2. METHODS...... 27 2.1. Relationship between bone finds and number of burials ...... 27 2.2. Estimating minimum and probable number of buried individuals ...... 27 2.2.1. West Estonian graves ...... 27 2.2.2. South-eastern Estonian graves ...... 28 2.3. Proportion of determinable cremated bone units in grave ...... 29 2.4. Colouration of cremated bones as an indicator of cremation temperature ...... 29 2.5. The fracture pattern of cremated bones ...... 30 2.6. Fragmentation stage of the cremains ...... 31 2.7. Assessment of sex and age at death ...... 31 2.8. Demographic estimations...... 32 2.9. Radiocarbon dating of bone ...... 34 2.9.1. Radiocarbon dating of cremation burials ...... 34 2.9.2. Radiocarbon dating of inhumation burials ...... 34 2.10. Statistical analysis ...... 35 3. RESULTS...... 36 3.1. Cremated bone material ...... 36 3.1.1. Radiocarbon dates ...... 36 3.1.2. The fragmentation of cremated bones and proportion of determined bone units in graves as an indicator of burial practice ...... 36 3.1.3. The colouration of cremated bones as an indicator of pyre temperature ...... 40 3.1.4. Character of cremated bone material and its relation to burial practice ...... 41 3.2. Demographic figures and community size ...... 42 3.2.1. Number of burials, biological sex and age at death assessments ...... 42 3.2.2. Radiocarbon dates and time-span of grave usage ...... 44 3.2.3. Demographic figures: mortality, fertility and community size ...... 46 3.3. Temporal and cultural differences in middle and late Iron Age west Estonian and south-eastern Estonian burial customs ...... 50 3.3.1. The grave types, dates and burial practice ...... 50 3.3.2. The influence of the funeral pyre on human remains ...... 51 3.3.3. Other agents responsible for bone fragmentation in recovered cremation burial deposits ...... 52

5 3.3.4. The character of cremated bones and burial customs in Iron Age Estonia ...... 54 3.3.5. Pyre sites ...... 58 3.3.6. Cemetery and community size ...... 59 3.3.7. Burial customs for children and infants ...... 60 CONCLUSIONS ...... 65 REFERENCES ...... 68 Appendix 1. Table 6. Community size calculations. Models ...... 81 Appendix 2. Measurements of cremated bone fragments. Tables 7.1.-7.2...... 82 Appendix 3. Lifetables non-corrected. Tables 8.1.-8.6...... 83 Appendix 4. Life tables corrected after Rösing and Jankauskas (1997). Tables 9.1.-9.6. ... 85 Appendix 5. Poanse demographic calculations. Tables 10.1.-10.4 ...... 87 PUBLICATIONS ...... 89 I. Some remarks on Kaseküla stone-cist grave, Läänemaa, Estonia ...... 91 II. Siksälä Kirikumägi: evidence of a new grave form of South-Eastern Estonia? ...... 101 III. Observations on Estonian Iron Age cremations ...... 119 IV. The demography of Iron Age graves in Estonia ...... 139 KOKKUVÕTE ...... 157 ELULOOKIRJELDUS ...... 160 CURRICULUM VITAE ...... 160

6 LIST OF PUBLICATIONS

This dissertation is based on four publications, which are referred to in the analytical overview by Roman numerals. I. Allmäe, R. (2010). Some remarks on Kaseküla stone-cist grave, Läänemaa, Estonia. Fennoscandia Archaeologica, XXVII, 45–52. II. Valk, H., Allmäe, R. (2010). Siksälä Kirikumägi: evidence of a new grave form of South-Eastern Estonia? Estonian Journal of Archaeology, 14(1), 40–55. III. Allmäe, R. (2013). Observations on Estonian Iron Age cremations. Archaeologia Baltica, 19, 31–47. IV. Allmäe, R. (2014). The demography of Iron Age graves in Estonia. Lietuvos Archeologija, 40, 103–120. The author’s responsibility for publication II was analysis of the osteological material and drawing conclusions based on cremated human remains, the preparation and writing of publication II was shared with the co-author.

List of other related publications Mägi, M., Allmäe, R., Maldre, L. (1998). Viking Age graveyard at Piila, Saaremaa. Archeological Fieldwork in Estonia 1997 / Arheoloogilised välitööd Eestis 1997, 99–116. Allmäe, R. (2003). Läänemaa 5.–13. sajandi kalmete antropoloogiline aines. In: M. Mandel, Läänemaa 5.–13. sajandi kalmed. Lisa 1. Tallinn: Eesti Ajaloomuuseum, 243–262. Allmäe, R., Maldre, L. (2005). Rõsna-Saare I kääbaskalmistu – esialgseid osteoloogilisi andmeid. In: Ü. Tamla, ed., Setumaa kogumik, 3: Uurimusi Setumaa loodusest, ajaloost ja folkloristikast. Tallinn: Ajaloo Instituut, 121–137. Allmäe, R. (2006). Grave 2 of Maidla – the burial site of a single family. Estonian Journal of Archaeology, 10(1), 3–23. Allmäe, R., Aun, M., Maldre, L. (2007). Cremations of the culture of long barrows in Northern Setumaa in the second half of the first millenium. Preliminary results. Humanbiologia Budapestinensis, 30, 113–122. Aun, M., Allmäe, R., Maldre, L. (2008). Pikk-kääbaste tähendusest (Rõsna küla kääbaskalmistute materjali alusel). In: Ü. Tamla, ed., Setumaa kogumik, 4. Uurimusi Setumaa loodusest, ajaloost ja folkloristikast. Tallinn: Tallinna Ülikooli Ajaloo Instituut, 269–290. Allmäe, R. (2008). Role of fire in burial customs. On the basis of two Iron Age burial places in Estonia. In: AEA 2008 Annual Conference, The consequences of Fire, 12.–14. September 2008, Århus, Denmark. Denmark: Moesgård Museum, 17–19. Valk, H., Allmäe, R. (2009). Põletusmatused Siksälä Kerigumäel. In: M. Aun, ed., Setomaa, 2. Vanem ajalugu muinasajast kuni 1920. aastani. Tartu: Eesti Rahva Muuseum, 387.

7 Allmäe, R., Aun, M., Maldre, L. (2009). Rõsna-Saarõ I kääbaskalmistu. In: M. Aun, ed., Setomaa, 2. Vanem ajalugu muinasajast kuni 1920. aastani. Tartu: Eesti Rahva Muuseum, 88–94. Mandel, M., Allmäe, R. (2009). Ergebnisse der archäologischen Ausgrabungen in Uugla. Archeological Fieldwork in Estonia / Arheoloogilised välitööd Eestis 2008, 114–123. Mandel, M., Allmäe, R. (2013). Forschungs- und Herrichtungsarbeiten im Gräberfeld zu Maidla. Archaeological Fieldworks in Estonia / Arheoloogilised välitööd Eestis 2012, 281– 288. Allmäe, R. (2014). Rõsna kääbaskalmistud Põhja-Setumaal: põletusmatuste uuringutulemused. In: Ü. Tamla, V. Lang, eds., Ajast ja ruumist. Uurimusi Mare Auna auks. Muinasaja teadus, 25. Tallinn; Tartu: Tallinna Ülikooli Ajaloo Instituut, Tartu Ülikooli ajaloo ja arheoloogia instituut, 39–50. Mandel, M., Allmäe, R., Maldre, L. (2015). Eine weitere Überraschung vom Gräberfeld zu Maidla. Archeological Fieldwork in Estonia / Arheoloogilised välitööd Eestis 2014, 107– 112. Аллмяэ Р., Аун М., Малдре Л. (2007). Предварительные результаты изучения остеогического материала курганных могильников Рысна-Сааре I и II в Северной Сетумаа (Юго-Восточная Эстония). In: Археология и история Пскова и Псковской земли. Семинар имени академика В. В. Седова. Материалы LII заседания, посвященного памяти профессора А. Р. Артемьева. Псков: Институт археологии РАН, 298–310. Аллмяэ Р., Аун М., Малдре Л. (2008). К вопросу о значении длинных курганов (по археологическим и остеологическим данным). In: Археология и история Пскова и Псковской земли. Семинар имени академика В. В. Седова. Материалы LIII заседания. Псков: Институт археологии РАН, 303–312.

8 ABSTRACT

Several Estonian Middle and Late Iron Age graves with cremations and inhumations from west and south-eastern Estonia were investigated in the period 1997 to 2011. During the research period various descriptive and metric data on cremated bone materials from west and south-eastern Estonia were recorded. The first aim of the study was to systematise and interpret the data collected on descriptive and metric features of cremated bone material in order to provide some generalisations on Estonian cremations. A comparative study of graves on the basis of the minimum number of buried individuals and the number of determined bone finds in graves, as well as bone fragmentation and colouration, is presented. Some conclusions on possible temporal changes and cultural differences in burial practices are made on the basis of these characteristics. Radiocarbon dating by the Accelerator Mass Spectrometry (AMS) method using burnt and inhumed human bones from six investigated graves was conducted in order to specify the usage time of the graves. The second aim of the study was to estimate the number of burials in graves, the age at death and sex of the deceased, and to model the demographic figures for some Estonian Iron Age communities. Three graves from south-eastern Estonia and two from west Estonia were suitable for palaeodemographic analysis. The five burial grounds studied were exploited by communities of 4–9 people usually, which corresponds to one family or household. The Rõsna-Saare I barrow cemetery in south-eastern Estonia indicated the possibility of a somewhat larger community, which may mean an extended family, larger household, or usage by two nuclear families. The crude death rate estimated on the bases of juvenility indices varied at Rõsna village between 55.1‰ and 60.0‰ (58.5‰ on average) during the Middle Iron Age. The estimated crude death rate was somewhat lower (38.9‰) in Late Iron Age Maidla, and extremely high (92,1‰) in Middle Iron Age Maidla, indicating an unsustainable community The third aim of the study was to assess possible regional, cultural and temporal differences in the handling of human remains in connection with burial customs in west and south-eastern Estonia. During the Middle Iron age in south-eastern sand barrow cemeteries only cremations were detected, although cremation was practiced with child or infant deaths. In west Estonian stone graves during the Middle and Late Iron Age, inhumation and cremation were both practiced; the burial practice of children and infants is ambiguous, as their cremated remains are rarely found in graves.

Key words: cremation, Iron Age, bone fragment size, radiocarbon dates, cremated bone, Estonia, burial custom, palaeodemography.

9 PREFACE

“Curiosity killed the cat… Satisfaction brought it back.” English proverb

I would say the first half of this phrase has always been true. The investigation of cremated human remains came out of pure curiosity. How much information can we acquire from burned bones in graves? Is an investigation even possible? What are the results? I should admit that there we moments when I felt very much like that curious cat. Today I know that to understand the message present in burnt bones takes time, a lot of time; in order to reach even general conclusions, add some more time. I hope one day the second half of the proverb will also come true. Cremated human remains are common finds in Estonian Iron Age graves, and studies of the artefacts, grave construction and funerary practice surrounding Iron Age burials using cremations have been conducted for decades and are numerous in Estonian archaeological research. Despite of the enormous amount of cremains collected from archeologically investigated burial places, materials studied osteologically are scarce in Estonia. There are various reasons for this. The first is the number of osteologists in Estonia. Researchers in the field of osteoarchaeology are very few in comparison with the number of active archaeologists. Therefore the new generation of anthropologists and osteoarchaeologists is warmly welcomed. Secondly, analysing cremated human and animal remains is time- and labour-intensive; the results of the work are seldom sensational. The determination of species, assessment of age and sex of cremated humans, measurement, counting the bones, assessment of colour and fracture pattern of bones and description of composition of bone assemblages, etc. – these are processes that take time, and sometimes several iterations. The third set of reasons would be methodological. The results of osteological analysis depend on excavation technique and need fair archaeological context, thus good collaboration with the archaeologists who excavate the burial place is mandatory. If you can participate in the archaeological fieldwork by yourself, this is even better. Osteological methodology is also important and depends on material. Methods that work for one material do not necessarily work for another. In the case of scattered cremations, the results of analysis might have little to say if the osteologist does not synthesise the results (for example, using established criteria to distinguish plausible burials in a grave) in order to facilitate conclusions. However, despite all of this, curiosity can do more than kill the cat. It can force you to find relations, reason and patterns, even if these are hidden to begin with.

10 ACKNOWLEDGEMENTS

I have had very good company on this road, and I hope it will last. Archaeozoologist Liina Maldre is the colleague I want to thank most. The time we spent together working with cremated bone materials from the graves of Läänemaa (west Estonia) and northern Setumaa (south-eastern Estonia) is precious. I am grateful to my tutor and dear colleague Leiu Heapost for her kind and subtle teaching through the years at the Institute of History. I am thankful to my tutor Marika Mägi for valuable comments on my papers and writings. Without the guidance and leadership of archaeologists Mare Aun and Mati Mandel the puzzle of cremations would have not been solved for these cemeteries: thank you. I am grateful to Rimantas Jankauskas, from Vilnius University, for enlightenment in the field of human cremation, and for encouraging discussion and for giving valuable advice. I’m very grateful to my reviewers, Professor Elisabeth Iregren and Dr. Laurynas Kurila; their comments and advice have been extremely helpful. There are so many colleagues and friends to whom I am thankful for inspiration, encouragement, discussion, support, and jokes; however, you know it already. My family – when I had good times, when I had hard times, you were always on my side. It kept me going.

11 INTRODUCTION

In Estonia cremation burials appear in the Late Bronze Age (1100–500 BC), for example in stone-cist graves and ship graves, however inhumation is still characteristic for the period (Lang, 2007a: 154; Jaanits et al, 1982: 150–151). Cremation burials are occasionally found beneath Late Bronze Age cists and Early Iron Age (500 BC–450 AD) tarand graves (Laul, 2001: 27–32; Lang, 2007a: 217– 218). In south-eastern Estonia, including Setumaa, the tradition to bury cremated human remains in pit graves also appeared in Bronze Age and lasted during the Pre- Roman period (500 BC–50 AD) and Roman Iron Age (50–450 AD), and even up to medieval times (Laul, 2001: 27–31; Kiristaja, 2009: 48–54; Lillak, 2009: 82; Aun, 2009: 83). During the Early Iron Age, cremations appear in cairn graves and are found occasionally in many Pre-Roman early tarand, where they appear with inhumations (Lang, 2006: 65–67, 2007a: 170–180; Kalman, 2000a, 2000b, 2000c, 2000d; Kivirüüt, 2014). In Roman Iron Age tarand graves cremation as well inhumation was practiced (Lang, 2007a: 191–205; Mägi, 2005a, 2005b); however, cremation was the prevailing burial practice during the Roman Iron Age, for example in tarand graves in south-eastern Estonia (Laul, 2001: 195–197; Lang, 2007a: 198– 201, 2007b: 144). Roman Iron Age (50 AD–450 AD) burial sites are not found in continental west Estonia (Lang, 2007a: 202; Mandel, 2000: 108, 2003: 11). At the beginning of the Middle Iron Age (450–800 AD) burial sites, for example stone graves without formal structure like Maidla I, Lihula and Ehmja ‘Varetemägi’, appear in Läänemaa, west Estonia; in these graves cremations as well inhumations are found (Mandel, 2003: 27–42, 128; Tvauri, 2012: 257). As well as underground cremation burial, the stone grave without formal structure is the most common grave type during the Late Iron Age (800–1200 AD) in west Estonia (Mandel, 2003: 170; Mägi, 2004; Tvauri, 2012: 258–268). In south-eastern and eastern Estonia sand barrows with cremation burials appear at the beginning of the Middle Iron Age. Cremation barrows are attributed to the Culture of Long Barrows, and are most numerous in the villages Laossina and Rõsna, in northern Setomaa, on the western shore of Lake Peipsi (Aun, 1992, 2009: 70–116; Tvauri, 2012: 269). The present study mainly concerns human skeletal remains collected from Middle and Late Iron Age burial grounds in south-eastern and west Estonia, and is the first attempt to generalise the results of research conducted during the 1997–2011 period. Human skeletal remains from the sand barrows of northern Setumaa (south- eastern Estonia) and from stone graves of west Estonia were the core of the research. Additionally, one Late Bronze Age stone cist grave in west Estonia was investigated to shed light on the burial customs relating to children. One Early Iron Age tarand grave and one Late Iron Age flat-ground cemetery in south-eastern were studied to find possible differences in the character of skeletal materials collected from graves.The inhumated and cremated human skeletal remains from ten graves from West Estonia (2008 bone assemblages) and five graves (548 bone

12 assemblages) from South-eastern Estonia were analysed, the results were systematised and interpreted in order describe burial sites and burial practise, and to find some temporal and/or regional differences in burial custom. Firstly the character of cremated bone assemblages was described. The colouration of bones was assessed in order estimate the cremation temperature and burial practice. To estimate the fragmentation stage of cremated bones in graves the longest tubular bone fragment and biggest cranial vault fragment were measured in bone units collected from graves. For this purpose the percentage of bone units where at least one determinable cremated human bone fragment was recovered, was also calculated. Secondly, the minimum, and the plausible, number of buried individuals, their biological age at death and their sex were estimated for graves in order to describe skeletal populations buried and to model the living communities who used the graves. Radiocarbon dating (AMS method) was used to specify the time-span of grave usage and to find out possible re-use of graves. The third aim of the research was to find possible temporal, regional and/or sociocultural differences in burial practice. For this purpose, the character of the skeletal remains and burial practice was analysed and interpreted. Special attention has been paid to the burial practice of children, and to the peculiarities of cremated bone units collected from possible pyre sites. The possibility to detect the ritual crushing of bones within the framework of the burial customs of Middle and Late Iron Age communities is discussed in light of recent research into cremated bones in neighbouring countries. In Estonian archaeological collections the amount of collected cremains is significant; however materials studied osteologically are scarce. The present study of osteological materials from Middle and Late Iron Age graves from west and south-eastern Estonia aims to fill this gap to some extent.

THE STUDY OF CREMATED HUMAN REMAINS IN NORTHERN EUROPE In discussions with my colleagues from countries around the Baltic Sea I have realised that most of the work done with cremated osteological materials is still reports in archives, very few of which have been published in recent years. Despite this I have found that various research has been published on cremated bones in Northern Europe, usually based on extensive material. Very intriguing aspects of cremation are studied, and rapidly developing scientific methods allow the study of the osteological materials in quite new ways. This will hopefully reclaim the cremation grave as a source of archaeological and osteological research.Cremated human and animal remains were deemed worthless in archaeology because of the fragmentation of bones and the little information they carried. The importance and value of cremated human remains is recognised mainly due to research of Swedish professor Nils-Gustaf Gejvall. In his osteological works Niels-Gustaf Gejvall demonstrated that cremations are a valuable source of

13 demographic estimates, and of cultural studies, of past societies (1948, 1959, 1961, 1963, and 1981). Elisabeth Iregren, a student of professor Gejvall, investigated cremated human and animal bones from the Iron to Viking Age at Varby-Varberg near Stockholm (Iregren, 1972). The methods Iregren used in her dissertation were mainly worked out by professor Gejvall (1948, 1963), although Iregren made several important conclusions about the composition and size of the human population and about the pathologies evidenced on bones; the composition of animal remains in graves; the age composition of domesticated animals; and how human and animal remains were burnt. She also made important observations concerning the representation of different skeletal elements in cremation layers and urn graves and established criteria by which to define pyre sites (Iregren, 1972). Later on, to mention only few, Iregren (1983, 1997) analysed and interpreted cremated osteological (human and animal) materials from several Swedish Iron Age graves, as well as making important conclusions on the burial customs of the Iron Age, especially on the basis of the accompanying animals in cremation graves. Pirjo Lahtiperä, also a student of professor Gejvall, who worked in cooperation with Unto Salo on Finnish Iron Age graves at the Valley of Kokemäe River, investigated cremated human bone material from 16 burial sites and determined the age at death and/or sex of 50 individuals. She also described the bone concentrations she studied, measured the long bone fragments and described bone colour (Lahtiperä, 1970: 199–219). Per Holck, Norwegian researcher investigated eastern Norwegian cremation graves from the Bronze to Late Iron Age (1986). With his study “Cremated Bones: A Medical-anthropological Study of Archaeological Material on Cremation Burials” Holck has achieved significant results relating to prehistoric cremation practice, pyre temperature and the peculiarity of burials, especially child burials, as well as discussing different aging and sexing methods, pathologies and the occurrence of animal bones in graves. Probably the most valuable part of Holck’s study is the proposed classification using grades of cremation (Holck, 1997/2008: 90–114). Holck has revised his study twice: in 1997 the analysis of Danish cremation graves were added, while in 2008 there were only minor changes (Holck, 1997/2008). Berit Sigvallius (1994) analysed Swedish Iron Age cremations from North Spanga in her doctoral dissertation “Funeral Pyres”. Apart from comprehensive analysis of human and animal remains she also made some cremation experiments with different animals, pyre constructions and different species of timber. She concluded that experienced people conducted the cremation process and that burial rituals included crushing the bones because the bones do not fragment to such high degree only through fire (Sigvallius, 1994: 31–33). The idea of ritually crushing cremated bones before burial is also suggested by Anders Kaliff (Kaliff 1992, 1997; Kaliff, Oestigaard, 2004). Sigvallius (1994, 118–120) also found out that in North Spanga the proportions of clean bones in graves diminishes and proportions of sooty bones increases from The Early Iron Age to the Late Iron Age, but both may occur in

14 parallel from the Migration Period to the Viking Age. For demographic calculations Sigvallius suggested only models taking into account only adults, because children’s cremains are seldom found in Iron Age graves. Caroline Arcini (2005) has found that the osteological composition of Swedish cremation graves is various, with especially highly variable presence of the petrous part of the temporal bone. The presence of petrous parts is frequent in compact burial assemblages in some kind of container (61–94%), and is very low (10–15%) in other burials (scattered cremations, bone layers, cremation pits, etc.). Arcini (2005) also proposed criteria to distinguish a burial place from the pyre site: the pyre site is usually a sooty patch with some charcoal and soot and tiny bone fragments; the absence of the petrous part of temporal bone is common. In Finland published osteological studies based on cremation graves are scarce. The first thoroughly investigated and published cremation cemetery in Finland is Vainionmäki A (Purhonen, 1996). Vainionmäki A is a cremation cemetery under level ground, where single and scattered human remains were found. Individual burials were not usually discernible at Vainionmäki A, due to the collective nature of most burials (Purhonen, 1996: 120). Tarja Formisto (1996) investigated the cremated bones at Vainionmäki A grave in order to estimate the number of burials in graves, the sex and the age at death of individuals to ascertain the composition of animal species and described bone material in general. Formisto (1996: 86) argues that in some cases the burials were carried out at pyre sites, because the bone fragments are often sooty. Formisto (1996: 81–87) found that the size of cremated bone fragments varied from 0.5–8 cm, and were usually 1–3 cm long. She argues that cremated bone fragments are of the same size in all cremation graves, and that it is very unlikely that bones were crushed before burial. The second published study on cremated bone material of Finnish cremation graves concerns osteological material from Rikala under the level-ground cremation cemetery at Salo (Mäntyla-Asplund, Storå, 2010). The archaeological and osteological study, which includes radiocarbon dating of cremated bone, demonstrate the quite complicated nature of Rikala cemetery and was based on scattered human remains of 14 individuals across multiple periods (1st–7th cc. AD). The authors emphasise the importance of radiocarbon dates and osteological analysis in the interpretation of cremation graves. In the 1990s in the Baltic States, cremations were mainly investigated to ascertain the number of burials in the grave, their age and sex composition and the demographic characteristics of the living population. Rimantas Jankauskas has studied 1st–7th century inhumation and cremation graves at Marvele (Jankauskas, Urbanavicius, 1998; Jankauskas, 2002, 2009), while Laurynas Kurila (2009) has investigated east Lithuanian barrow cemeteries, finding that the results of osteological ageing and sexing of cremains was an important contribution to the study of burial customs as well as to demographic calculations. Recently Laurynas Kurila has published an impressive series of papers on cremation graves (Kurila, 2015a, 2015b, 2015c). Kurila (2015a) analysed the accuracy of sex estimations

15 based on cremated remains and found the accuracy of sexing female graves is 85.5%, although the accuracy for male graves is much lower at 52.5%. In his subsequent work, Kurila (2015b) analysed the male and female graves in order to identify social classes in east Lithuanian Iron Age communities. He found connections with wealth and the military dimension of the male graves, and for this reason sharper stratification among males than among females. Recently (Kurila, 2015c) a comparative study of inhumation and cremation burials was conducted in order to assess the reliability of cremated human remains as a source of demographic modelling. The author found one significant limitation of cremated burials – the deficit of identified sub-adults, which in most cases causes biased demographic profiles. However, in his study Kurila found that many demographic figures display a rather close match between inhumation and cremation samples, emphasising the value of macroscopic study of cremated bones. In Latvia, Gunita Zariņa (2006, 2009) has also included the analyses of cremation burials in the large-scale study of Latvian demography from the Mesolithic to the Early Modern period. Zariņa also detected dramatic fluctuation in population size because of economic and political processes in the Laukskola Liv community of the 10th–13th centuries (Zariņa, 2006, 2009: 180–184). In Estonia investigation of Middle and Iron Age cremation graves has mainly focused on the number of individuals buried in the grave, their age and sex composition; in addition the work ascertained the number of animal species in the grave and the relation of these remains to the burials. The practice of studying cremation graves is somewhat different in Estonia in comparison with other Baltic States, and with Scandinavia or Finland. In Estonia, this material has been studied by a close cooperation of archaeologists, anthropologists and archaeozoologists, and the results of this work published in many scientific papers and books (Mägi et al, 1998; Mandel, 2003; Allmäe, 2003; Maldre, 2003; Allmäe, Maldre, 2005; Allmäe et al, 2007a, 2007b, 2008; Aun et al, 2008; Mandel, Allmäe, 2009; Valk, Allmäe, 2009, 2010; Mandel et al, 2015). The same practice is followed in north-east Russia, where the graves of the Long Barrow Culture are being studied (Khvoshchinskaya, 2004). In recent years in Scandinavia research into cremated human bones has focused more on methodological issues, experiments and uses of new technologies, for example stereology, computed tomography and radiography of burial vessels, together with osteological study of cremated bones (Harvig et al, 2012, 2014a; Jaeger, Johansen, 2013; Harvig, Lynnerup, 2013). Experimental archaeology has demonstrated that the cremation of piglets about the size a human infant leaves a remarkable number of burnt bones. According to Jaeger and Johansen (2013) the reason why infant remains are seldom found amongst cremations is probably caused by other factors, such as burial custom or a separate burial place/form for infants. Harvig and colleagues (2012) developed a methodology to study container burials. Combining osteological and other scientific methods, the results of investigations are more reliable as well easier to conduct. Harvig and Lynnerup (2013) have found

16 that in Danish Late Bronze Age urns the burials often contain the remains of entire cremated individuals, and the ritual crushing of cremains before placement into burial urns was not observable. Based on osetological study, CT scans, and estimation of volume and weight of cremains in urns, the authors argue that under- representation of archaeologically recovered human remains may be to a large extent due to taphonomic processes. The authors also proposed methodological suggestions to estimate the fragmentation of cremated human remains: the Fragmentation Index, i.e. the weight in grams (g) of cremated bones divided by volume (cm3), a highly valuable method when container burial is under study. Lately Harvig and colleagues (2014a) have conducted a comparative study of osteological materials from Late Bronze Age urn burials and Early Iron pit graves at Fraugde, Denmark, in order to follow the peculiarities of osteological material in different grave types and across changes in burial practice. They found important differences in cremation temperature and handling of the bone after cremation, the fragmentation stage and the composition of cremains for different Late Bronze Age and Early Iron Age burials/graves. According to the authors the cremation pits of the Early Iron Age are intentional secondary deposits, as pyre debris is rarely found in this pit. During last decades cremated bones in graves have gained new value. Rapid development of scientific methods in archaeology enable us to radiocarbon date graves on the basis of burnt bones (Lanting et al, 2001) to seek the origin of buried people with the help of isotopic studies (Harvig et al, 2014b) or to look for ancient DNA in cremation graves (Pusch et al, 2000). The experiment to artificially enrich dental enamel and calcinated boned with strontium demonstrated that calcinated bone is more resistant to post-mortem exchange than tooth enamel (Snoeck et al, 2015), which makes calcinated bones a promising source for strontium isotope studies. The possibility to date cremated bones has opened the door to a re- evaluation of chronologies of Bronze and Iron Age graves in Europe (De Mulder et al, 2007, 2009, 2014; Mäntyla-Asplund, Storå, 2010; Allmäe, 2013, 2014a). The reliability of AMS dating of cremated bone has been questioned, and experiments have been conducted to ensure this method is being used successfully (Naysmiths et al, 2007). The studies to evaluate the effect of funeral pyre materials on radiocarbon dates of cremated bone are important in the field of archaeological chronology (Olsen et al, 2013, Hüls et al, 2010). Therefore, cremated human remains are a valuable source in the study of past populations and societies, including demography, and in the study of these cultures’ mortuary customs.

THE OSTEOLOGICAL RESEARCH ON CREMATED HUMAN REMAINS IN ESTONIA Cremated human remains from Estonian Iron Age graves have been studied since the 1990s. The analyses of cremated bones are generally macroscopic, i.e. their

17 main target has been the identification of the number of people buried, their biological sex and age at death (Kalling, 1993; Mägi et al, 1998; Kalman, 2000b; Allmäe, 2003). A somewhat different approach to investigating cremation cemeteries has been adopted by archaeologist Marge Konsa, who conducts intrasite spatial analysis of cemeteries. She has studied artefact distribution patterns and cremated bones in 10th–13th cc. Madi cemetery (south Estonia) with dispersed cremation burials (Konsa, Engbring, 2011; Konsa, 2013). Konsa (2013) found that the correlation between general distribution of bones and artefacts was very weak. Apparently cremated bones were distinguished from the rest of the pyre remains and treated separately from grave goods. The first known study of cremated human bones in Estonia was conducted by Ken Kalling (1993), who analysed cremated human remains (cremains) from the Viimsi I and II tarand graves, dated to 350–500 AD (Lang, 1993: 55, 2007a: 133). Anu Kivirüüt (2011, 2014) has revisited the material twice. In 1997 Marika Mägi excavated the Viking Age graves at Piila. The cremated bone material was analysed by author and archaeozoologist Liina Maldre (Mägi et al, 1998). Two conclusions were made on the basis of osteological analysis. First, the cremation temperature was high: the bones were all white, calcinated and highly fragmented. The second conclusion was that the majority of identified bones belonged to animals; human bones were rarely identified in the grave. Maldre began collaborating with archaeologist Mati Mandel in 1998. Mandel has excavated west Estonian cremations since the 1970s (Mandel, 2003). The cremated and non-cremated human and animal bones of eight 5th–13th century stone graves were analysed, including analysis of long-bone fragment size and some observations on bone colour (Allmäe, 2003; Maldre, 2003). In his study of the 5th– 13th century stone graves of Läänemaa (west Estonia) Mandel (2003) combined the archaeological and osteological data. Jonathan Kalman (2000a, 2000b, 2000c) has made determinations of cremated bones occurring in Early Iron Age tarand graves. In 2014 Anu Kivirüüt (2014) has revisited the cremated (and inhumated) osteological materials of Tandemäe and Viimsi I and II tarand graves, analysed before by Kalman (2000b) and Kalling (1993), respectively. Kivirüüt (2014) used osteological and spatial analysis methods to receive new data concerning the Iron Age tarand graves in Northern Estonia. The author estimated the minimum number of burials and the cremation grade of burials; she also presented demographic data and described peculiarities of burials, including handling of the corpses. New information was presented regarding the burial rituals and grave layout of Iron Age society in northern Estonia. Since 2004 cremains from Long Barrow Culture (6th–10th cc.) sand barrows in south-eastern Estonia have been analysed. Mare Aun excavated the sand barrow cemeteries between 1970 and 1990 (Аun, 1992). The combined study of the result of the analysis of burial archaeology and osteology are published in several papers,

18 including analysis of bone fragment sizes (Allmäe, Maldre, 2005; Allmäe et al, 2007a, 2009; Aun et al, 2008). I found out that cremation as a burial ritual was common in adult as well child deaths in the area of Long Barrow Culture in south- eastern Estonia. Moreover, age segregation was found, as one sand barrow (no 9) contained only sub-adult burials. In addition, an important discovery was that no sacrificed animals were found in child burials. Leiu Heapost (2007) identified cremations at Kalmetemägi (Siksälä in south-eastern Estonia). In 2008 the cremains of the Uugla III stone grave (11th–13th centuries) in western Estonia were analysed and published (Mandel, Allmäe, 2009). In 2009 the cremains of the Kirikumägi flat ground cemetery (990–1160 AD) in south-eastern Estonia were analysed (Valk, Allmäe, 2009). In 2010 the archaeological and anthropological study of Kirikumägi flat ground cemetery was published, including analysis of the colour and size of bone fragments (Valk, Allmäe, 2010). In both cases new grave form was discovered – flat-ground burials of cremated human remains. During the research 1997–2011 various descriptive and metric data on cremated bone materials from Läänemaa (west Estonia) and southern Estonia were collected by the author. The results of analysis were systematised and generalising observations on cremated bone materials and burial practice were presented by author (Allmäe, 2013, 2014a, 2014b). In these papers, conclusions on temporal changes in cremation burial practise were made based on the correlations between the radiocarbon dates and the bone fragmentation stage in different graves. A comparative study of the graves on the basis of other characteristics, for example the minimum number of buried individuals, the number of determined bone finds among the total number of bone finds, and the colour of the cremated bone fragments in graves is also presented. The colour of the bones has a descriptive value in this study, and is only briefly discussed. Between 2012 and 2015 small excavations at Maidla’s second stone-grave were conducted by Mati Mandel to investigate the whole grave area; during these years some new inhumation and cremation burials were discovered (Mandel, Allmäe, 2013; Mandel et al, 2015).

THE HISTORY OF PALAEODEMOGRAPHIC INVESTIGATIONS IN ESTONIA Palaeodemographic studies are not very numerous in Estonia, as there are some requirements for material and data, for example the whole burial place should be excavated, migration should be excluded, skeletal material should be complete, accurate data concerning burial chronology and origin of series should be available, etc. (Acsádi, Nemeskéri, 1970; Alesan et al, 1999).

19 The first attempts to analyse the demography of Estonian ancient populations were made on the basis of archaeological data (Ligi, 1989; Lang, Ligi, 1991; Lang, 1996). These authors used the type and number of artefacts and number of bone assemblages in graves to calculate the size of the community that used the burial place, and made assumptions on population density and population size based on the burial grounds in different Estonian districts. The first palaeodemographic calculations based on osteological research into prehistoric graves were conducted for the Roman Iron Age tarand grave Viimsi II in northern Estonia (Lang, 1993), and for the Pre-Roman tarand in Poanse in western Estonia (Kalman, 2000a). Ken Kalling (1995, 1997) performed the first known study of palaeodemography based on archaeoanthropological materials from Medieval and Early Modern Tartu. Kalling analysed the skeletal population buried in Tartu’s Jaani Church during the 13th–14th centuries and proposed a demographic model for population. The skeletal sample from 14th–18th century Tääksi village cemetery (southern Estonia) was demographically analysed by the author (Allmäe, 1998); the material contained some cremations from the 15th–16th centuries (Sokolovski, 1990; Allmäe, 1998). The second grave of Maidla in western Estonia contained 10th–11th century cremations and inhumations from the 12th–13th centuries (Mandel, 2003). Demographic analysis based on these skeletal materials was conducted by the author of the present paper (Allmäe, 2006). Leiu Heapost (2007) has analysed the demography of the 11th–15th centuries Kalmetemägi in Siksälä in south-eastern Estonia. That study concerns mainly inhumation burials, but some cremations were also included in the calculations. Heapost analysed temporal changes in mortality figures, estimated life expectancies for different age groups and calculated population size. Between 1997 and 2011, several Middle (450–800 AD) and Late (800–1250 AD) Iron Age cremation graves in western and south-eastern Estonia were analysed by the author. During the research period the estimations of the minimum, and plausible, numbers of buried individuals in graves, the assessment of their sex, and their age at death was conducted. Amongst the material studied there are five burial places that are a good source for demographic calculations. New radiocarbon dating methodology (Lanting et al, 2001) enables one to date the cremated bone material, helping to date cremation graves where artefacts are rare and therefore archaeochronological dating is difficult. The radiocarbon dates (AMS method) of burnt human bones from six investigated graves are presented in this paper. The goal of these analyses was to specify the time span over which the communities used these graves. In 2010 the search for missing infants and children in Late Iron Age graves in Läänemaa (west Estonia) concluded with a publication in which radiocarbon dating and osteological analysis of inhumed infant burials from Kaseküla Late Bronze Age stone-cist grave confirmed the reuse of the grave 2000 years later for infants (Allmäe, 2010).

20 The results of this research enabled the calculation of demographic figures, which in turn enabled conclusions to be drawn on prehistoric burial practices and usage period of graves for Middle and Late Iron Age graves in west and south-eastern Estonia (Allmäe, 2014 a, 2014b).

21 AIMS

The overall objective of the present study is to summarise the research on Middle and Late Iron Age human cremations in Läänemaa (west Estonia) and south-eastern Estonia. The research had three main aims. The first aim was to summarise the results of descriptive and metric analyses of cremated bone material from west and south-Eastern Estonia in order to find links and patterns between burial custom and character of cremated bone material. The second aim was to establish demographic figures and models to characterise the Iron Age communities who used the graves during certain periods. The third aim was to find out if any regional or socio-cultural peculiarities, or temporal changes, can be identified in Iron Age burial customs.

22 1. MATERIAL

During the research period a total of 2556 bone finds or bone units from west Estonia and south-eastern Estonia were analysed. The material mainly comes from Middle (450–800 AD) and Late Iron (800–1250 AD) graves. The exceptions are Põlgaste tarand grave, which is from the Early Iron Age or Roman Iron Age (500 BC–450 AD) and Kaseküla stone-cist grave, from the Late Bronze Age (1100–500 BC). The material analysed consists mainly of cremated human bones, and to some extent inhumations. The minor part is formed of inhumed burials in situ, mainly recovered from Maidla stone graves and from Kaseküla stone-cist grave. It should be noted that cremated human remains in Estonian deposits are not washed and sieved. The material is often packed with some soil and pyre debris. Most materials are collected from excavations without sieving; the sieving of soil has only become common practice in Estonian archaeology quite recently, i.e. since the 1990s.

1.1. WEST ESTONIA The list of analysed materials from West Estonia (Läänemaa) is as follows (Figure 1):  Kaseküla stone-cist grave in western Estonia was archaeologically investigated by Mati Mandel in 1973. Stone-cist graves are above-ground structures that have one or several cists in the middle and which are enclosed by one or several circular stonewalls, filled with soil and stones and covered with a stone heap (Lang, 2007a: 147). Inside the central stone-cist commingled unburnt human bones and a fragment of a bronze razor were found; outside (north) of the stone encirclement remains of another human burial were found (Mandel, 1973, 1975: 74). The fragmentary and commingled bone material from Kaseküla stone-cist grave area has been gathered by squares (2 m x 2m); the number of collected bone finds is 81, the material is deposited in the Estonian History Museum. The bone material was analysed by the author twice in 2009, as a discrepancy occurred between the former result (Kalman, 2000d) and the new analysis conducted by the author. A lot of bone material was also taken for analyses by another researcher between the two periods of analysis. It must be noted that in the 1970s the soil removed from archaeological objects was not sieved before disposal, which explains why most of the (expected) human teeth are absent from the deposited material. The osteological material of Kaseküla stone-cist grave has been popular among researchers and has been analysed by several (Kalman, 2000d; Allmäe, 2010). According to Margot Laneman (2012), the material was not analysed by osteologist Martin Malve, although Laneman’s paper is based on the osteolgical remarks of the latter. Mandel (1975) dated the burial site to the second half of the first millennium BC; later on Valter Lang (1996: 297) dated it to the Late Bronze Age (1100–500 BC). Lang also

23 suggested the possibility of secondary burials in the grave (Lang, 1996: 297), and later research confirmed these suggestions (Allmäe, 2010; Laneman, 2012).  Lihula stone grave from 5th–7th centuries AD was investigated by Mati Mandel in 1974 (Mandel, 2003: 27–30). 144 bone finds were collected by 2 m x 2 m plots from the grave area; the material contained commingled cremated and non- cremated human remains; 21 burials were recorded (Mandel, 2003; Allmäe, 2003).  Keskvere II underground cremation burial from the 7th–8th centuries was investigated by Mati Mandel in 2001 (2003: 103–104, 123–133); 10 bone finds were collected from a 4.3 m x 3 m plot. The material contained cremated and non-cremated bone fragments (Mandel, 2003: 104; Allmäe, 2003: 245–246).  The stone graves of Uugla (I, II, III) were investigated by Mati Mandel in 1977, 1981 and 2008, respectively; the osteological material (513 bone finds) was collected by 2 m x 2 m plots. The graves are dated 11th–13th centuries and contain only cremations (Mandel, 2003; Mandel, Allmäe, 2009). Ehmja stone grave from the 5th–7th centuries AD and from the 10th to 13th centuries AD was excavated by Mati Mandel between 1982 and 1991. 248 bone finds were collected from the plots of 2 m x 2 m. Eight burials pits were recorded, probably from the 5th–7th centuries. In addition, scattered cremation burials from the 10th– 13th centuries were recorded in the upper layers of the grave. Ehmja stone grave was repeatedly disturbed during prehistoric times. Mandel (2003) has suggested this happened as a result of the reuse of the cemetery. The grave contained cremated and uncremated human bones (Mandel, 2003; Allmäe, 2003).  The stone graves of Maidla I (5th–6th centuries AD) and II (10th–13th centuries AD) were excavated by Mati Mandel between 1983 and 1985, and between 1987 and 1990 (Mandel, 2003: 39). The first Maidla stone grave was mainly used during the 5th and 6th centuries AD, while some 11th–13th century artefacts were also found in the southern part of grave (Mandel, 2003: 42). The bone material consists of 268 finds, collected by 2 m x 2 m plots. The grave generally contained scattered cremation burials although some inhumations were also found (Mandel, 2003: 40–41; Allmäe, 2003). The second stone grave at Maidla was used mainly from the second half of the 10th century to the beginning of the 13th century AD (Mandel, 2003, 59). The bone material consists of 674 finds, collected by 2 m x 2 m plots. The grave contained inhumations as well scattered cremation burials (Mandel, 2003; Allmäe, 2003, 2006). In recent years new discoveries have been made at the Maidla grave field, for example a double inhumation burial was found in the periphery of the grave area in 2008 (Mandel, Allmäe, 2013), and some inhumations were recovered in 2014 (Mandel et al, 2015).  The Kirbla stone grave, dating from the 11th–12th centuries AD, was excavated by Mati Mandel in 1983 (Mandel, 2003: 90). 70 bone finds were collected by

24 2 m x 2 m squares. The human bone material consisted only of burnt bones (Mandel, 2003; Allmäe, 2003). The analysed bone material from west Estonia consists of 2008 bone finds.

Figure 1. Investigated Estonian Iron Age cemeteries.

1.2. SOUTH-EASTERN ESTONIA The list of analysed material from south-eastern Estonia is as follows (Figure1):  The Põlgaste tarand grave, dating to the 3rd–5th centuries AD (Early Iron Age) was excavated by Silvia Laul between 1970 and 1973 (Laul, 2001: 40–43). Tarand graves are burial places with characteristic quadrangular stone enclosures, commonly with a north–south orientation (Jaanits et al, 1982: 207; Lang, 2007a: 170, 192). The bone material from the Põlgaste grave was collected by 1m x 1m plots and is all cremated (Laul, 2001: 42). The cremains of the Põlgaste tarand grave were analysed in 2010–2011. Unfortunately the author could not find any whole-bone material excavated from the grave. The only bone finds available were collected from the western, earlier, part of the grave where bone material was slightly burned in comparison with that of the eastern tarands (Laul, 2001: 42, 196). The analysed part of the cremains consists of 59 bone units.

25  Suure-Rõsna, Rõsna-Saare I and II sand-barrow cemeteries from the second half of the 1st millenium AD were excavated by Mare Aun during the 1980s. Rõsna- Saare I cemetery consisted of ten barrows, Rõsna-Saare II cemetery of eleven barrows and Suure-Rõsna ten barrows, of which six were excavated (Aun, 1992: 100–105). The barrow cemeteries consist of rounded and long barrows of piled sand, with various burial customs apparent; the cremated bones were commonly buried as quite compact assemblages under and into the barrows (Aun, 1992, 2005; Aun et al, 2008). The human bone material collected from the barrows is all cremated (Allmäe, Maldre, 2005; Allmäe et al, 2007a; Allmäe, 2013). The total number of analysed bone units from the barrow cemeteries is 420.  Kirikumägi flat-ground cemetery from the late 10th–12th century AD was excavated by Heiki Valk in 2003–2004 and 2007. The scattered cremation burials were detected in an underground level. 69 bone assemblages were collected. The bones were mainly cremated, with the exception of some unburnt teeth and temporal bones from a 3- or 4-year-old child (Valk, Allmäe, 2010). The analysed bone materials from graves of south-eastern Estonia consist of 548 bone finds.

26 2. METHODS

The present study is based on macroscopic investigation of cremated bones.

2.1. RELATIONSHIP BETWEEN BONE FINDS AND NUMBER OF BURIALS Usually the cremated bone fragments are collected from the burial site and packed as assemblages or units. All these assemblages/units are numbered. Some archaeologists call these units bone finds; some have argued that this packed unit of bone fragments is not an archaeological find, but an assemblage of bones. It does not matter what we call these packed bone units, it is important to define what the unit, find or assemblage we study is. To begin with we have to realise that the bone assemblage/unit/find is not automatically one burial. This is especially true for burial places where cremated bones are scattered, and somewhat true for other types of grave where the burial technique is not recognised on site. How well we understand the burial technique later depends on excavation methodology: how the material was excavated, collected, packed and described on site: put simply, how well archaeologists understand the bone material on site and how osteologists understand the excavation situation during the process of analysis. However, one packed and numbered collection of bone fragments gathered during archaeological excavations is a good unit to study the peculiarity of the material, the fragmentations and dispersal of cremated bones in different graves.

2.2. ESTIMATING MINIMUM AND PROBABLE NUMBER OF BURIED INDIVIDUALS

2.2.1. West Estonian graves

2.2.1.1. West Estonian cremation graves It is known, that some parts of skeleton are more resistant to heat-induced changes and preserve better among cremated bones (Holck, 1997/2008). Thus, the minimum number of buried individuals (MNI) among the examined cremains is very often determined according to the recurrent bone fragment method. Several different bone fragments were counted during analyses, but in most cases the pars petrosa of os temporale was the best unit for estimating the MNI. In some cases, where recurrent fragments did not occur, the individuals were distinguished on the basis of estimates of biological age (for example, if in a bone assemblage the cranial vault fragments of an adult and unfused epiphysis of a long bone, or the crowns of deciduous teeth were found, etc.). The probable number of individuals (PNI) was estimated on the basis of fragments (20 or more) of cranial vault combined with at least one determined fragment of

27 human skeleton. The cranium was preferred because the fragments are easily determinable amongst cremated bones, and most likely full corpses were cremated. To estimate the PNI distances between bone units were also taken into account. The PNI is an important unit for graves with scattered or disturbed cremations (Maidla I and II graves, Ehmja, Kirbla, Uugla I–III, Lihula). On one hand, in most cases the bone fragments are collected according to quite large plots, i.e. 2 m x 2 m; on the other hand, we do not know how much material was carried from pyre to the burial place. The material was analysed and the PNI was determined according to the criteria described above.

2.2.1.2. Kaseküla stone-cist grave with inhumations The method of recurrent bone fragments (for example, pars petrosa of the temporal bone, the diaphyses of infant long bones, etc.) was used to estimate the minimum number inhumations in the grave.

2.2.2. South-eastern Estonian graves

2.2.2.1. South-eastern sand barrow cemeteries Material from barrows has also been collected as bone assemblages. The assemblages here are much more representative examples and perceived as concrete burials. The MNI was estimated for every barrow in all analysed sand-barrow cemeteries. Numerous different recurrent bone fragments were recorded; however, here also the pars petrosa of os temporale, as the most frequently found element, was the best unit for calculating the MNI. Unfortunately the number of cranial fragments was not counted because at the beginning it seemed that the bone assemblages were well-defined units, i.e. archaeologically distinguishable burials. The collected and deposited units were named main and additional assemblages. It was not always possible to decide if the additional bone assemblage(s) were part of the main one or not, and sometimes it seemed that the numbered main bone assemblages were not equal to one burial. The discrepancy between the MNI and archaeologically determined number of burials/assemblages is obvious; the osteologically estimated MNI is usually smaller than the archaeologically estimated. There are two reasons for this. First, the criteria for burial are not clearly distinguishable during excavation. It is nearly impossible to decide which of two main assemblages an additional assemblage might belong to, or if it is a separate burial. Secondly, as is often the case we do not know how much of the cremains were collected from the pyre and buried in the barrows. Finally the PNI was determined taking into account the distances of collected bone units, the MNI, the colour of the cremains, the size and composition of the collected bone unit, and the results of age and sex determinations. In comparison with former published studies (Allmäe, Maldre, 2005; Allmäe et al, 2007a; Aun et al, 2008) in

28 later publications (Allmäe, 2013, 2014a) some recalculations of PNI and MNI for sand-barrow cemeteries have been performed.

2.2.2.2. Tarand-grave and flat-ground cemeteries in south-eastern Estonia In both cases the osteological material was collected by square plots. In Kirikumägi flat-ground cemetery at Siksälä the bone assemblages are smaller than in Põlgaste tarand grave. The MNI and PNI were estimated for both cemeteries. Unfortunately the materials represent only part of these burial places: Kirikumägi has only been partially excavated and from Põlgaste tarand grave some of the bone material went missing from the depository.

2.3. PROPORTION OF DETERMINABLE CREMATED BONE UNITS IN GRAVE The number of bone units, where determination of at least one human bone fragment was possible, was calculated for the ten graves under study. The hypothesis is that this characteristic describes the dispersal of bones – the high number of units with non-determined bone fragments should define a burial place with scattered cremation burials.

2.4. COLOURATION OF CREMATED BONES AS AN INDICATOR OF CREMATION TEMPERATURE Bone colour was not recorded according to the accepted colour standards, as the material has been analysed over quite a long period (1997–2011). During that time the methodology of recording bone material changed to some extent. However, the general notes and impressions of bone material are briefly discussed, as some pattern was observable. The changes in bone colour due to cremation is discussed in the light of works by Shipman et al (1984), McCutcheon (1992), Holck (1997/2008), Walker et al (2008) and Wahl (2008). “Charred” bones occur in cremations when oxygen flow and heat were insufficient for proper combustion (Holck, 1997/2008, Walker et al, 2008, Wahl, 2008). The bone fragments indicating this appear as blackish or dark brownish amongst the cremains: according to McCutcheon (1992) these specimens have been heated to 340 ºC; according to Holck (1997/2008) grade 1 combustion (200–400 ºC), to Wahl (2008) grade 2 (300–400 ºC), and stage II/III (285–525 ºC) according Shipman et al (1984). The appearance of black/dark brownish bones was recorded in the studied bone materials. The percentage of bone units with brownish/blackish bone fragments is presented as a percentage of the total number of bone units within the studied material. A grey colour appears at the temperatures of 550 ºC to 650 ºC according to Wahl (2008) and Walker et al (2008); Shipman et al (1984) have described light grey

29 with secondary colours of brown and light brownish grey from 525 ºC to 645 ºC; McCutcheon (1992) observed light brownish grey at temperature up to 600 ºC; Holck (1997/2008) has determined dark and light grey colours at temperatures of 400 ºC– 800 ºC. The grey colour is attributed to bones that are not completely burned. At high temperatures, of 800 ºC or more, the bones become calcinated and a white colour is prevalent, according to Holck (1997/2008). McCutcheon (1992) and Wahl (2008) argue that the white colour appears at the somewhat lower temperature of 650–700 ºC. According to Shipman et al (1984) the predominant colour of bone is a neutral white with blue-grey or light grey at the temperatures up to 940 ºC. When bones are heated in excess of 940 ºC, the specimens are neutral white with some medium grey and reddish-yellow (Shipman et al, 1984). ‘Sooty’ bones occur in burials where the bones are buried with pyre remains. This means that the cremated bone fragments and other pyre remains are collected from the cremation site and buried in the grave, or that the cremation place and the grave are the same, or that we have found the actual cremations place itself. Sooty bones were found in all the above-mentioned cases. When clean bones occur in a grave we may conclude that the cremation pyre is located away from the grave (Sigvallius, 1994). Clean burned bones in graves give a clue about a burial technique in which bone fragments are carefully gathered from the cremation site without pyre remains. The occurrence of sooty bones was recorded for the graves under study here.

2.5. THE FRACTURE PATTERN OF CREMATED BONES The alteration of bones due to heat has been studied for decades and is different if they are dry bones, fresh/green/de-fleshed bones, or fleshed bones/corpses. The thermal fractures and alteration characteristic to the cremation of corpses are warping, transverse, especially curved transverse, fractures, and longitudinal fractures, splintering and delaminating (often seen on cranial bone), patina, etc. (Holck, 1997/2008; Buikstra, Swegle, 1989; Whyte, 2001; Mayne Correia, 1997; Symes et al, 2008; Musgrave et al, 2010). The results of experimental studies done to distinguish the condition of human remains before cremation have been ambiguous, even contradictory (Mayne Correia, 1997; Fairgive, 2008: 50–52; Gonçalves, 2012: 21–24; Larsson, Nilsson Stutz, 2014). For example recently Gonçalves and colleagues (2011) have also found occasional curved fractures and warping on cremated dry bone and suggest it is related to the preservation of collagen in bone rather than being an indicator of fleshed or recently de-fleshed human remains being burned. The warping of dry bones has also been observed by Buikstra and Swegle (1989). However, the cracks and fractures that appear on bones vary depending on whether the soft tissue is present or not during cremation (Larsson, Nilsson Stutz, 2014). The presence of deep multiple lateral cracks, deep checking, curved fractures, and also uneven colouration of bones are commonly the indicators of burning fleshed corpses (Larsson, Nilsson Stutz, 2014). Hence, when

30 all or some of these features are absent or occur rarely there is a possibility that dry bones or bones with a low collagen content were cremated. The heat-induced alterations of bones were recorded macroscopically to estimate if the corpses were burned. However, the inconsistent observation of various patterns has to be admitted, as the pattern was not recorded for every bone unit during the research on the first archaeological samples.

2.6. FRAGMENTATION STAGE OF THE CREMAINS The most reasonable way to estimate the fragmentation of cremains is to measure the biggest bone fragment(s) in one packed unit or bone assemblage. The bone fragment size was measured for all analysed materials. The bone fragment size is post-excavational, though it must be noted that the material was not sieved or washed during excavation, or after. The maximum length of the long-bone fragment was measured for every packed bone unit for the ten burial places under study. In addition, the biggest fragment of cranial vault was measured for ten burial places. Two measurements were taken for cranial vault fragments: the maximum length and the transversal length/width. The metric characteristic of the cranial vault size is expressed as cross multiplication of these two measurements. A sliding calliper was used to measure the bone fragments.

2.7. ASSESSMENT OF SEX AND AGE AT DEATH The sex and age at death of the deceased were determined according to common osteological standards (Miles, 1963; Workshop, 1980; Brothwell, 1981; Buikstra, Ubelaker, 1994; White, Folkens, 2000; Bass, 2005; Mays, 2006). No archaeological gender assessments were available during biological sex and age assessment. Specific to cremated bone material is that most of the age and sex related morphological features on the post-cranial skeleton are not observable. To assess biological sex of the cremated adult individuals the morphological traits of the occipital bones (external occipital protuberance, nuchal lines), the temporal bones (mastoid process, temporal line, zygomatic process), the frontal bones (superciliary ridge, supraorbital margin, temporal ridge, glabella), the zygomatic bones (frontal process), and mandible (condylar process, mandibular ramus and angle, mental tubercules and protuberance) were used. In some cases morphological traits on the post-cranial skeleton (for example muscle attachment areas of the long bones) were used as additional criteria; general robustness of skeletal elements was also observed. To assess age at death of cremated individuals the suture closure of cranial bones, the union of epiphyses and dental development were used. If applicable, other criteria were used to estimate or specify the age at death: the morphology of the cranial vault (Gejvall in Sigvallius, 1994) and the morphology of tooth roots – the

31 roots of older individuals become more rounded due to the deposition of cementum. Hypercementosis is quite common in older individuals (Acsádi, Nemeskéri, 1970; Soames, Southam, 1993), which can also add data to ageing attempts. Age-related pathologies on cremated bones, for example osteoarthritis or osteophytosis on vertebrae or anywhere on skeletal elements, were also used. The ageing of sub- adults is more reliable due to stages of epiphyseal fusion and dental development. It must be emphasised that the age and sex determinations based on cremated human remains are less reliable than those based on inhumations. This is because of the incompleteness of human remains in cremations due to high fragmentation of bones; thus only a few skeletal elements are available for identification. In the case of Kaseküla stone-cist inhumations the maximum lengths of the long bone diaphyses of infants were measured to estimate the age at death. The values were compared to the standard correlation between age estimates and maximum lengths of infant long bone diaphyses (Lovejoy et al, 1990) and to the same correlations compiled on the basis of Italian and Estonian archaeological material (Facchini, Veschi, 2004; Allmäe, 1998).

2.8. DEMOGRAPHIC ESTIMATIONS Demographic estimations were performed using several methods. In our model we assume that the population is stationary and that birth and death rates are equal (growth = 0). The reason is that Estonian populations under study are all too small- scale to model positive or negative natural increase. Firstly, the method of life tables proposed by G. Acsádi and J. Nemeskéri (1970) 0 was used to estimate the life expectancy at birth (e 0). Here the natural data received from skeletal samples are used. Secondly, the life tables were corrected according to F. W. Rösing and R. Jankauskas (1997), where the proportion of small children (0–4 years) in the population was increased to 45% of the total skeletal population under the study. This means that 45% of the population died before they reached the age of 5. J. P. Bocquet and C. L. Masset (Bocquet, Masset, 1977; Bocquet-Appel, Masset, 1982) established the third model used here to estimate the demographic figures for past populations. The ratio of sub-adults to adults or the juvenility indices was also calculated for every population under the study: number of children deceased between 5 and 14 number of adults deceased at 20 and later in short: D5–14/D20+ The model was worked out to get over the problem of small children (0–4 years) often being under-represented in the graves. The juvenility index allows us to estimate demographic parameters without the bias due to infant under-

32 representation in osteological collections. Additionally it allows for control of systematic bias in the calculation of adult age distribution. The formula (Bocquet, Masset, 1977; Bocquet-Appel, Masset, 1982) to estimate newborn life expectancy from the juvenility index is following:

0 e 0=78.721*log10√1/x–3.384±1.503

where x=D5–14/D20+ 0 The newborn life expectancy, crude death rate (1/e 0), and the size of living populations was calculated according to all three models. In these models we assume that the population is stationary, i.e. birth and death rates are equal (growth = 0). The size of the living population was calculated according to the formula proposed by D. Ubelaker (1989): 0 P=N*e 0/T where P = population size, N = number of burials at cemetery, 0 e 0 = life expectancy at birth (in years), T = the time span over which the burial place was used (in years). In our model we made assumptions concerning the time-span of the cemetery’s use. To model the population sizes for Rõsna cemeteries the estimated usage period of the grave (T) is 150 years and for Maidla first grave 75 years. The exception is the second Maidla grave, where 250 years of grave usage was established by archaeologist Mati Mandel (2003), and is used in the presented calculations. However, the usage period of the grave is an estimate for archaeological populations, as we are never able to know the exact beginning or abandonment time of a particular burial site. An overview of how community size depends on the proposed usage period of the grave is presented in Appendix 1. The calculated community size declines if we extend the grave usage period, and grows if we reduce it. While the archaeological/typological date of a grave might be for example 400 years, we cannot be certain that the grave was used continuously for the whole period. Therefore for calculations we should pick a model that indicates the sustainability of the community for some time in particular conditions (the population is stationary, the birth and death rates are equal). Reproduction was estimated according to two different models. In the first model fertility rate and the number of female offspring born per woman (GRR) was estimated from the juvenility indices of Boquet-Appel and Masset (1982) and calibrated from R. McCaa (1998, 2000), then the total number of offspring per woman (TFR) was calculated: GRRx2.05=TFR.

33 Henneberg (1975) established the second model we used to estimate reproduction from archaeological human remains. The approach is somewhat different in comparison with conventional palaeodemography. Henneberg (1975) combines demographic (mortality structure) as well as biological characteristics of human fertility to construct reproduction model for human palaeopopulations. To estimate the reproduction rate of the population the following definitions and calculations are used: the potential gross reproduction rate (Rpot) or the average number of births per couple of adult individuals during their lifetime; the net reproduction rate (Ro) or the average number of adult descendants per adult individual of parental generation, and the absolute number of offspring born per average adult couple (C). To calculate the last figure we need a hypothetical value Uc (the total number of births achievable throughout the full reproductive period). The closest to reality, Uc is around eight (Acsádi, Nemeskéri, 1970), but it could be lower or higher. In the present study the number of offspring born per average couple (C) was calculated using a Uc value of 7.45 (Lorimer, 1954 in Henneberg, 1975). The masculinity index for every population was calculated by dividing the number of males by the number of females in the population.

2.9. RADIOCARBON DATING OF BONE

2.9.1. Radiocarbon dating of cremation burials Nine samples of cremated human bones from six graves were dated. The radiocarbon dating of cremated bone fragments from Maidla I (Hela-2403) and Maidla II (Hela-1958) stone graves, Rõsna-Saare I (Hela-1959 Hela-1960) and Rõsna-Saare II (Hela-1961) barrow cemeteries, Suure-Rõsna (Hela-1962) barrow cemetery and Põlgaste tarand grave (Hela-2404, Hela-2405) was performed in 2009 and 2010 at the Dating Laboratory of the Finnish Museum of Natural History, University of Helsinki. The 14C concentration was measured using the AMS method. The results were calibrated according to the Intcal09 curve (Reimer et al, 2009) and Oxcal 4.1 software (Bronk Ramsey, 2009).

2.9.2. Radiocarbon dating of inhumation burials Three samples of human bones from the Kaseküla stone-cist grave (Poz-32412, Poz-32413, Poz-32414) were radiocarbon dated in Poznan Radiocarbon Laboratory using the AMS method; calibration was made by OxCaL 3.10 software Bronk Ramsey (2009), atmospheric data from Reimer et al (2009). Three pieces of bone were removed from deposited material for radiocarbon dating in summer 2009: a fragment of the cranial vault of an adult from the central stone- cist (find No 29), the left femur of an infant in the central stone-cist (find No 29) and the petrous part of the left temporal bone of an infant from the encirclement area (find No 40).

34 One non-cremated bone fragment (Hela-1919) from Maidla II was also dated at the Dating Laboratory of the Finnish Museum of Natural History, University of Helsinki. The 14C concentration was measured using the AMS method. The results were calibrated according to the Intcal09 curve (Reimer et al, 2009) and Oxcal 4.1 software (Bronk Ramsey, 2009).

2.10. STATISTICAL ANALYSIS The idea was to compare different cremation graves on the basis of correlations between observed characteristics. The average size of long-bone fragments, the proportion of determined bone finds amongst material, and radiocarbon dates of the graves should indicate differences in burial practice. It should be possible to observe the changes in burial practice over time and also the cultural and geographical differences this reveals. R 2.15.2 was used for statistical analysis.

35 3. RESULTS

3.1. CREMATED BONE MATERIAL

3.1.1. Radiocarbon dates Radiocarbon dates (AMS method) of cremated bones from six of the graves under study were conducted to specify the time period when the grave was in use (Table 1, Allmäe, 2013). The main purpose of this analysis was to find out if there is any correlation between the characteristics of bone material and the time period when the grave was in use. The relative dates of stone graves at Maidla (Mandel, 2003) are as follows: the Maidla I grave dates to the 5th–6th centuries AD and Maidla II to the 10th–13th centuries AD. The 14C dating of cremated bone material confirms the gap, but indicates usage of the Maidla I grave somewhat earlier. The south-eastern Estonian sand barrow graves studied are dated to the 6th–8th centuries AD (Aun, 1992). The radiocarbon dates indicate usage of sand barrow cemeteries somewhat earlier (Table 1). The tarand grave of Põlgaste was dated to the 3rd–5th centuries AD (Laul, 2001: 42–43). The AMS dates of cremated bones from the earlier part of grave here also show somewhat earlier usage of the grave (Table 1). The tarand grave of Põlgaste was archaeochronologically dated to the 3rd–5th centuries (Laul, 2001: 27, 40–41). Radiocarbon dates of cremated bones from the older part of the grave indicate somewhat earlier establishment of the burial place – at the end of the Pre-Roman or the beginning of the Roman Iron Age (Table 1). The flat ground cemetery Kirikumägi at Siksälä was radiocarbon dated on the basis of charcoal amongst the cremains (Valk, Allmäe, 2010) producing grave dates from the end of 10th to the beginning of the 12th century (Table 2).

3.1.2. The fragmentation of cremated bones and proportion of determined bone units in graves as an indicator of burial practice The fragmentation of cremated bones and proportion of determined bone units were estimated for ten graves (Table 2, Appendix 2: Tables 7.1.–7.2.). The proportion of bone finds determined is quite small for west Estonian Late Iron Age stone graves with scattered cremations. The new form of cremation grave, Iron Age flat ground burials was detected in Kirikumägi, Siksälä, in south-east Estonia (Valk, Allmäe, 2009, 2010). Here, as in west Estonian stone graves, the cremains were scattered and the number of bone units determined was relatively small. The cremains in sand barrows from the Long Barrow Culture in south-east Estonia are buried under barrows or in barrows, and the buried bone assemblages are mainly quite large and compact here (Aun, 1992, 2005; Aun et al, 2008). The Maidla first grave in west Estonia is contemporaneous with south-eastern sand barrows, in addition to

36 Table 1. Radiocarbon dates (AMS method) of cremated bones

Sample Location in Radiocarbon Calibrated date Grave Specimen Analysed Lab number d13C(‰) number grave determination bone fragment (95.4%) (68.2%) Square plot Left ulna 258 (11.9%) 297AD Maidla I MI-240 104F Adult HELA- 2403 -22.4 1675 +-30 BP 341 – 413 AD proximal end 321 (83.5%) 427 AD stone grave Square plot Maidla II MII-180 96/96 n-o Adult Cranial vault HELA-1958 -22.2 1000 ± 30 BP 980-1160 AD 990 – 1120 AD stone grave Square plot 40 BC (91.5%) 88 AD Põlgaste Po-2 8/b-c Adult Cranial vault HELA- 2404 -23.9 1565 ± 35 BP 5 – 74 AD 103 AD (4.4%)122 AD tarand grave Square plot 20 BC (5.4%)12 BC Põlgaste Po-3 10/f Adult Cranial vault HELA- 2405 -20.9 1595 ± 35 BP 45 BC – 80AD 1 BC (62.8%) 65AD tarand grave Long barrow Rõsna-Saare no 7. - Adult Axis HELA-1959 -27.2 1620 ± 35 BP 410 – 570 AD 430 – 540 AD I 1st bone set under barrow. Rounded barrow no 9. Rõsna-Saare 420 AD (29.3%) 470 AD RSI-157 3rd bone set in Sub-adult Long bone HELA-1960 -26.7 1535 ± 35 BP 390 – 550 AD I 480 AD (38.9%) 540 AD the center of barrow Long barrow Rõsna-Saare no 7. 3rd set in 390 AD (36.3%) 460 AD RSII-86 Adult Cranial vault HELA-1961 -24.4 1959 ±30 BP 340 – 540 AD II the center of 480 AD (31.9%) 540 AD barrow Rounded barrow no 6. 430 AD (32.2%) 490 AD Suure-Rõsna SR-40 7th bone set in Adult Cranial vault HELA-1962 -22.5 1977+-30 BP 420 – 600 AD 510 AD (1.5%) 520 AD the center 530 AD (34.5%) 580 AD barrow.

37 Table 2. Descriptive and metric features of cremated bones in the analysed graves.

West Estonian graves South-eastern Estonian graves Name of grave Rõsna-Saare Rõsna-Saare Kirbla Ehmja Maidla I Maidla II Uugla III Põlgaste Suure Rõsna Kirikumägi I II Number of studied 70 248 268 674 326 59 159 97 164 69 bone units Grave type Stone-grave Stone-grave Stone-grave Stone-grave Stone-grave Tarand grave Sand barrow Sand barrow Sand barrow Flat-ground Archaeochronological 5–7 5–6 11-12 10–13 11–13 3–5 6–8 7–8 6–8 10–12 dates* (cc.) 11–12 (11–13) Radiocarbon dates 40BC–122AD 410–570AD – – 258–427AD 980–1060AD – 340–540AD 420–600AD 990–116AD** (95.4%) probability 45BC–80AD 390–550AD % of determined 13 17.2 41.9 29.7 18.1 81.7 70.6 66.6 55.1 25.6 bone units Minimum number of 1 5 17 32 7 18 34 22 26 4 cremation burials Number of measured long bone fragments 51 151 205 541 316 54 106 61 140 29 per grave Longest bone fr. in bone unit; ave per 3.2 2.2 4.3 3.2 2.7 5.7 4.7 4.9 4.3 3 grave (cm) Number of measured 27 75 139 248 192 47 65 40 84 15 cranial fragments Biggest cranial fragment in bone 3.4 2.7 7.4 5.0 3.2 12.9 8.6 8.9 7.3 4.6 unit. ave per grave LxW (cm) % of bone units with 2.6 1.8 9.2 9.7 5.8 3.4 11.4 9.4 4.8 2.6 "charred" bones

*from Mandel, 2003; Laul, 2001; Aun, 1992; Valk, Allmäe 2010; ** radiocarbon date from Valk, Allmäe, 2010

Figure 2. Correlation between radiocarbon date, bone fragmentation and proportion of determined bone units in grave. which the fragmentation of cremains and number of determined bone finds is also comparable. The Põlgaste tarand grave is the earliest, with the proportion of determined bone finds being highest and bone fragmentation lowest. It seems that burial practices are quite easily distinguishable through the percentage of bone units determined and fragmentation of cremains.The proportion of determined bone units is varies between the 10th–13th century graves and the 1st BC–8th AD century graves, the latter with higher proportions (Figure 2). The graves from earlier graves also had less bone fragmentation. Both measurements (cranial and long-bone fragment) decline with time – the earlier graves with burials of compact bone assemblages show a lower amount of fragmentation than the later ones. This also confirms that the custom of scattered cremation burials is distinguishable on the basis of higher bone fragmentation. The size of average bone fragments in graves diminishes with time – in Late Iron Age graves the bone fragmentation is higher. Thus, we may argue that Late Iron Age cremation can be distinguished from earlier sand barrow cemeteries by this descriptive feature.

39 3.1.3. The colouration of cremated bones as an indicator of pyre temperature The colouration of cremated bones was assessed for eight graves. The cremains in West Estonian stone graves (Maidla I–II, Uugla III) were commonly pale, often with white and grey colours dominating, sometimes with brownish and blue hues, thus the temperature of the cremation pyre would have been at least moderate. For example in the Uugla III stone grave, 87.1% of bone finds are described as white, with the blue hue also found very frequently here. For the Maidla II grave, Mandel (2003: 42–45, 59) has described the bone fragments as heavily burned and observed occasional melting of bronze artefacts in the grave. The Maidla I stone grave contained slightly and heavily burned bones (Mandel, 2003: 30–40) and bone fragments here were also slightly bigger in comparison with other studied western Estonian graves (Allmäe, 2003: 250–251, 2013). The bone fragments were mainly of a pale colour sometimes with a brownish hue. Wahl (2008) has observed white with brownish, greyish and ochre colours at 800 ºC and above, and argues that these secondary colours correspond to local soil conditions. The proportion of bone finds with a dark brown and/or black colour among west Estonian graves was more frequent in the Maidla stone graves (Table 2). Sooty bones were quite common in west Estonian stone graves too, as the cremation often took place in the grave area (Mandel, 2003: 153; Allmäe, 2003). The studied south-eastern Estonian graves with cremations are of varied construction: Põlgaste is a tarand grave, Kirikumägi is a flat ground cemetery and others are sand barrow cemeteries (Rõsna-Saare I–II, Suure-Rõsna). The prevailing colour of bones within the Põlgaste tarand grave material is pale with a light brown and occasional orange-red hue. Archaeologist Silvia Laul (2001: 196), who excavated Põlgaste tarand grave, has argued the bones have been slightly burned. The cremated bone material in Põlgaste grave included calcinated fragments and surprisingly uniform colouration. The latter is attributed to the cremation of dry bones; here calcinated bones are light brown or tan (Buikstra, Swegle, 1989: 249–252). Dark brown and black fragments were observed only very rarely among the material (Allmäe, 2013). In Kirikumägi flat ground cemetery mainly light coloured bone assemblages were found, with shades varying from light yellow greyish to light brownish; occasionally dark grey shades were also observed - the cremation temperature would have been al least 600 ºC, most likely above 800 ºC (Valk, Allmäe, 2010). More detailed observations were conducted to describe the colouration of the cremains in south-eastern Estonian sand barrows (Suure-Rõsna, Rõsna-Saare I and II). The prevailing colour of the bone units in sand barrows was white with a yellow hue (39.5% of bone units), with the next most common colour described as pale with a greyish beige or greyish yellow hue (37.1%), and brownish or dark grey colouration recorded for 10.2% of bone units. Dark brown and black colours (incompletely burned bone fragments) were recorded for 4.8–11.4% of bone units

40 (Table 2), while bone units buried with charcoal and soot (‘sooty bones’) formed 11% of buried bone assemblages (Allmäe, 2013). According to the observations on colouration of archaeological cremains from the Middle and Late Iron Age the cremation temperature would have been above 600 ºC, and most likely above 800 ºC (Valk, Allmäe, 2010; Allmäe, 2013). The occasional find of unburnt or charred fragments are not unusual among cremains (Table 2). In addition the grey colouration is common, showing incomplete burning of the organic part of the bone. The secondary surface colour, for example yellow or beige, or sooty bones, is caused by soil conditions or burial practice.

3.1.4. Character of cremated bone material and its relation to burial practice Nearly all bone materials showed fracture patterns and surface patterns characteristic to burning corpses, the only exception being the tarand grave at Põlgaste from the beginning of first millennium. Here the fracture pattern and colouration cremains could indicate possible secondary burial – the cremations of dry bones (Allmäe, 2013, 2014b). The colour of bones gives us a hint of the heat of the cremation pyre and also of the burial environment. White bones, sometimes with a yellow hue, are evidence of quite high pyre temperatures, definitely at least 800 ºC. White and pale bones were often found in west and south-eastern Estonian graves. A yellow hue on the surface of bones and in bone spongiosa is quite common for cremations above 800 ºC (Wahl, 2008). The greyish beige or greyish yellow colour of bone assemblages, which often occurred in south-eastern barrows, is much more difficult to interpret. The suggestion is that the bones are moderately burned, plausibly at temperatures between 600 and 800 ºC. Most likely, the beige colour is caused from the burial environment, while a greyish tone is partly caused by incomplete oxidation of fragments and partly by the burial environment. A grey and light grey colour is observed for a wide range of pyre temperatures, mainly beginning around 550–650 ºC. These colours were also found for example in Maidla I stone grave. The dark greyish-brown bone units are probably cremated at lower temperatures than 600 ºC (McCutcheon, 1992; Shipman et al, 1984; Holck, 1997/2008; Walker et al, 2008; Wahl, 2008). The dark greyish brown bone units were occasionally found from nearly all graves, for example in south-eastern sand barrows 10.2% of bone units were accounted in this group. These bone units are possibly the indication of unsuccessful cremation, something (for example weather) could have hindered the burning process. Bone units consisting of white fragments with a yellow hue were frequent in sand-barrows, suggesting that these bones were ‘clean’, not buried with charcoal or soot but carefully picked from the pyre remains before burial. For example, ‘clean’ bones form 30.2% of bones from Swedish Early Iron Age cremations and only 0.8% of Late Iron Age cremations (Sigvallius, 1994). On the other hand many (11.0%) of bone assemblages were buried in sand barrows with a remarkable amount of charcoal and soot – meaning with pyre remains. The occurrence of ‘clean and white’ and ‘greyish brown and sooty” bones in south-

41 eastern sand barrows demonstrates different contemporaneous burial practices for cremains. In western Estonian stone graves sooty bones were quite common as the cremation often took place in the grave area (Mandel, 2003: 153–159; Allmäe, 2003, 2013, 2014b).

3.2. DEMOGRAPHIC FIGURES AND COMMUNITY SIZE

3.2.1. Number of burials, biological sex and age at death assessments The number of analysed graves is fourteen; the main results are presented in Table 3. In many west Estonian graves (Maidla I and II, Ehmja, Lihula, Keskvere II) cremations as well as inhumations were detected (Allmäe, 2003, 2013; Mandel, 2003). In the studied south-eastern graves cremation was prevalent (Aun, 1992; Laul, 2001; Allmäe, 2013), except for the Kirikumägi flat-ground cemetery at Siksälä, where unburnt fragments of a skeleton of a 3–4-year old child were found (Valk, Allmäe, 2009, 2010). The minimum number of burials (MNI) and the plausible number of burials (PNI) differ in most cases, and the discrepancy is expectedly greater in larger graves. The biological sex of buried individuals was frequently not determinable (Table 3), especially for cremation burials. Worth mentioning from west Estonia are the graves at Maidla. In the second Maidla grave an estimation of biological sex was possible for only 18 of 42 cremated individuals ten female and eight male individuals were distinguished (Allmäe, 2003: 248). Four male and five female burials were distinguished amongst 17 cremation burials in the Maidla I grave (Allmäe, 2003: 245). In south-eastern graves only cremations were present, however the bone material had a much higher diagnostic value and the success rate in sex estimations was expectedly higher (Table 3). The proportion of sub-adults in graves is 50% or below (Table 3). The proportion has, of course, less importance in small graves (Uugla I–III, Kirbla), where only a few individuals are buried and/or cremated on the site (Allmäe, 2003, 2013, 2014a; Mandel, 2003). In partially investigated graves (Ehmja, Lihula, Keskvere, Põlgaste, Kirikumägi) the age structure can be distorted because only some of the osteological material was available for anthropological study. Therefore the proportion of sub-adults is significant in graves that have been thoroughly investigated: Maidla I, Maidla II, Rõsna-Saare I, Rõsna-Saare II and Suure-Rõsna. It seems that the proportion of sub-adults (aged under 15 years) is quite similar for these graves – from 34.4% to 40.0%. The only exception is the second Maidla grave, where the proportion of sub-adults is slightly lower – 29.7%. Even then the proportion of sub-adults could be overestimated here because it is highly probable that the infant burials, or at least some of them, are from the Medieval period (Allmäe, 2014a).

42 Table 3. Minimum and plausible number of burials; results of age and sex assessment. Name of the grave Maidla Maidla Keskvere Lihula Ehmja Uugla I Uugla II Uugla Kirbla Rõsna- Rõsna- Suure- Kiriku- Põlgaste I II III Saare I Saare II Rõsna mägi stone- stone- flat- stone- stone- stone- stone- stone- stone- sand sand sand flat- tarand Type of the grave grave grave ground grave grave grave grave grave grave barrow barrow barrow ground grave 5–6 5–7 Period * (cc) AD 10–13 7–8 5–7 11–13 12–13 11–13 11–12 6–8 7–8 6–8 10–12** 3-5 (11–13) 11–12 Number of bone units 268 674 10 144 248 157 30 326 70 159 97 164 69 59 Archaelogically determined nested ------65 53 41 - - bone units ** Minimu number of 19 51 1 10 5 4 1 7 1 34 22 26 4 18 individuals (MNI) Plausible number of 20 74 3 21 10 4 1 11 2 61 46 40 4 23 individuals (PNI) Cremations 17 42 3 10 5 4 1 11 2 61 46 40 3 23 Inhumations 3 32 - 9 5 ------1 - Subadults (0-15 y.) 8 22 1 4 5 0 0 1 1 21 18 15 1 4 Adults ( 15+ y.) 12 52 2 17 5 4 1 10 1 40 28 25 3 19 Males 5 13 - 2 2 - - 2 - 14 10 8 1 12 Females 5 16 - 4 - 1 1 3 1 15 8 8 1 6 Proportion of sex 0.50 0.39 n/a 0.29 0.2 0.25 1 0.45 1 0.48 0.39 0.4 0.5 0.78 estimations Masculinisation index 1.0 0.81 - 0.5 - - - 0.67 - 0.93 1.25 1.0 1.0 2.0 Undetermined sex 2 23 2 11 3 3 0 5 0 11 10 9 1 1 *archaeochronological dates from Aun, 1992; Laul, 2001; Mandel, 2003; Aun et al, 2008; ** radiocarbon date from Valk, Allmäe, 2010.

43 3.2.2. Radiocarbon dates and time-span of grave usage

The radiocarbon dates (AMS) of cremated and non-cremated bones from six investigated graves are presented in Table 4. The number of AMS dates is small, but do help to specify the time span over which the grave was in use. According to AMS dates the first stone-grave of Maidla is from the 3rd–5th centuries (archaeologically dated to 5th–6th cc.); the cremations in the second grave are from the 10th–12th centuries, as is already suggested by Mandel (2003). One infant inhumation from this grave was also radiocarbon dated, showing that apart from the 12th–13th century inhumations (Mandel, 2003), some of the infant inhumation(s) could be even be from Medieval times (Allmäe, 2014a, Table 4). The sand barrows from south-eastern Estonia have been radiocarbon dated to the 4th–6th centuries, probably indicating a somewhat earlier establishment of the graves in comparison with archaeochronological dating from 6th–8th cc. (Table 4; Allmäe, 2014a.) Radiocarbon dating of cremated bone possibly yielding a somewhat older age is a phenomenon supported by experimental evidence (Hüls et al, 2010; Olsen et al, 2013). Hüls and his coworkers (Hüls et al, 2010) found that depending on the cremation temperature, its duration, the composition of the burning atmosphere, and the composition/age of the fuel, an aging effect of 50–100 years may be possible. Therefore we cannot exclude the possibility that relatively old AMS dates are because of the burial techniques employed. On the other hand, there are available radiocarbon dates from charcoal amongst cremated bones collected from cremation burials found under the barrows of Rõsna-Saare I (Tln-285; 1825±80 BP) and in the Suure-Rõsna cemetery (Tln-735; 1920±35 BP). The charcoal beneath the first barrow of Rõsna-Saare I cemetery gave a calibrated date of 25–391 AD (95.4% probability), and Suure-Rõsna’s 6th barrow (95.4% probability) gave a date of 1– 221 AD (Tvauri, 2012: 272). However, the dates from cremated human bones (Table 1, 4) were not only collected from beneath the barrows, but also from the burials inside the barrows. The infant inhumations in Kaseküla stone-cist grave at Läänemaa were radiocarbon dated to specify when these burials took place, and if the infant burials are contemporaneous with Bronze Age burials or from later periods. The results of AMS dating showed that infants were buried during the Late Iron Age – 2000 years later than the adults in the stone-cists (Table 4; Allmäe, 2010). Radiocarbon dating (AMS) of burnt and unburnt bones in the graves is a good tool to specify when the burial place was used and when it was reused again after a time gap. The radiocarbon dates of cremated bones from five investigated graves were used to model the demographic figures and usage period of these graves (Table 5).

44 Table 4. Radiocarbon dates of cremated and inhumed human bones in graves. Archaeochro- Burial place Construction Bone sample Location in grave nological date Lab no BP 95.4 % 68.2 % of the grave* stone grave adult, left ulna Maidla I 104/F 5th – 6th cc. HELA-2403 1675 +-30 BP 258 – 427 AD 341-413 AD irregular prox., cremated stone grave adult, cranial Maidla II 96N–96O 10th – 13th cc. HELA-1958 1000 ± 30 BP 980 – 1160 AD 990-1120 AD irregular vault, cremated infant, os stone grave temporale, pars Maidla II 75S 12th – 13th cc HELA-1919 440 ± 30 BP 1410 – 1610 AD 1430-1465 AD irregular petrosa, uncremated Rõsna- 1st bone set under barrow no 7, adult, axis, Saare the barrow, 6th – 8th cc. HELA-1959 1565 ± 35 BP 410 – 570 AD 430-540 AD long cremated I cremated Rõsna- RSI-157 in the barrow no 9, subadult, long Saare center of barrow, 6th – 8th cc. HELA -1960 1595 ± 35 BP 390 – 550 AD 420 – 540 AD rounded bone fr., cremated I cremated RSII-86 in the Rõsna- barrow no 7, adult cranial vault, center of barrow, 7th – 8th cc. HELA-1961 1620 ± 35 BP 340 – 540 AD 390 – 540 AD Saare II long cremated cremated Suure- barrow no 6, adult cranial vault, SR-40 in the center 6th – 8th cc. HELA-1962 1535 ± 35 BP 420 – 600 AD 430 – 580 AD Rõsna rounded cremated of barrow, cremated adult cranial vault, Kaseküla stone cist No 29, in stone cist 1100–500 BC POZ-32412 2780 ± 30 BP 1010 – 840 BC 980 – 895 BC uncremated infant, left femur, Kaseküla stone cist No 29, in stone cist 1100–500 BC POZ-32413 1195 ± 30 BP 710 – 940 AD 760 – 900 AD uncremated infant, petrous part No 40, in stone Kaseküla stone cist of left temporal circle, square plot 1100 – 500 BC POZ-32414 920± 30 BP 1020 – 1190 AD 1040 – 1160 AD bone, uncremated 9/Q *Dates from Mandel 2003, Aun 1992, Lang 1996

45 3.2.3. Demographic figures: mortality, fertility and community size The modelled demographic characteristics for five Estonian graves under study are presented in Table 5; life tables with raw data are given in Appendix 3, Tables 8.1.– 8.6.; and life tables with corrected data in Appendix 4: Tables 9.1.–9.6. Newborn life expectancy, calculated according to non-corrected life tables (raw data), indicates the highest values for graves, and points to relatively low mortality. The data, corrected according to Rösing and Jankauskas (1997), in which the proportion of infants and small children (0–4 years) is increased to 45% of the total number of individuals, shows lower and probably more reliable values for newborn life expectancy (Table 5). The model based on the juvenile ratio (Bocquet, Masset, 1977; Bocquet-Appel, Masset, 1982) proposes the lowest newborn life expectancy. The juvenility indices D5–14/D20+ for sand barrow cemeteries vary from 0.28 to 0.31 (0.30 for the 0 summarised Rõsna sample), e 0 between 16.7 and 18.1 years (17.1 years for summarised sample) indicating mortality between 55.1‰ and 60.0‰ (58.5‰ for the summarised sample). These mortality values are much higher in comparison with those received from non-corrected life tables (31–38‰), and slightly higher in comparison with the mortality obtained from corrected life tables (45–54‰). The Maidla stone graves show higher variability in juvenility indices and newborn life expectancy (Table 5). The Maidla I grave indicates a very high crude death rate 0 (92.1‰) and extremely low newborn life expectancy (e 0=10.9), conditions under which a community is hardly sustainable. For the Maidla II grave the mortality rate 0 is the lowest (38.9‰), and life expectancy at birth (e 0=25.7) the highest in comparison with other communities under study. We should bear in mind that Maidla I is a very small grave (only 20 burials), and that in the Maidla II grave the older children (over 5 years) could also be underrepresented as well as infants as only two cremations of sub-adults were detected in the grave (Allmäe, 2003). The GRR and TFR fertility rates received from the juvenility indices are also presented in Table 5. According to this model, the lowest number of offspring is characteristic to women of Late Iron Age Maidla, and the highest to Middle Iron Age women at Suure-Rõsna. The number of offspring (TFR based on juvenility index model) increases with mortality, as expected, because higher infant mortality reduces the birth interval. When a nursing child dies, the natural sterility due to the lactation period is interrupted and conception happens earlier (Wood, 1990). Human reproductive strategies may change under the influence or interaction of several economic, social and biological factors. Fertility may increase, for example during war, famine and drought, although the mortality rate is high at the same time. This phenomenon is observable in the studied material: low newborn life expectancy is related to a higher number of offspring. The number of children born per woman was also calculated according to the model proposed by M. Henneberg (1975), showing more realistic values. The

46 cumulative number of deliveries per woman or per adult couple is from 3.88 to 5.06 in the communities under the study. The number of offspring varies from 3.82 to 5.06 for south-eastern barrow cemeteries, and between 3.88 and 4.8 for west Estonian stone-graves (Table 5). Ro or net reproduction rate indicates the replaceability of generations; Ro>1 for all observed communities suggests a positive increase. The sizes of living populations calculated according to different life expectancies indicate that in most cases one household or family used the graves under the study. The model also suggests that the Rõsna-Saare I cemetery might have been used by a somewhat larger household or extended family, or was used for somewhat longer period. The calculated size of living population is sensitive to the estimated time of grave use: if the time span diminishes the size of the calculated population will be larger, according to these models (Table 5; Appendix 1). The first Maidla stone grave in west Estonia was probably established at the end of the Roman Iron Age or at the beginning of the Middle Iron Age. One family at most 0 used the grave, and the crude death rate was very high (e 0=10.9). This meant that the community was probably not sustainable and perhaps inhabited the site for a short period only. The same pattern characterises the Pre-Roman Iron Age tarand graves at Poanse. Newborn life expectancy here is extremely low at 10.8 and 14.4 years, respectively (Appendix 1; Appendix 5: Tables 10.1.–10.4.). However, the crude death rates are very high and the calculated community sizes very small (2.9 people at Maidla I, 2.0 at Poanse I and 3.4 at Poanse II). The time-span over which these graves were used is 75, 250 and 150 years, respectively. It might be that the assumptions on time-span are wrong and the graves were used for a shorter period. The estimate of the population size is highly dependent on the term of grave use and of estimated life expectancy at birth. Our idea that graves, especially tarand graves, were used for several centuries could be somewhat overestimated. For example, if we diminish the term of grave use of Poanse I from 250 to 100 years, and of Poanse II from and 150 to 50 years, and use newborn life expectancy derived 0 0 from the juvenility ratio (e 0=10.8; e 0=14.4 years), we get another reality (Appendix 1). The size of community for Poanse I tarand grave is 5 and for Poanse II is 10. During the Late Iron Age a community of 7–10 individuals used the second Maidla grave in western Estonia – probably one family or household (Allmäe, 2006). The estimated family size is in accordance with the earlier results of various authors on the average size of the Estonian family in the 13th century and later periods (Blumfeldt, 1937; Ligi, 1961; Tarvel, 1972; Palli, 1996); Mägi (2002: 11, 74, 123) has argued that that for example stone graves on Saaremaa belonged to just one or two elite families. The present study also suggests that probably one family or household (8–9 individuals) had its own burial ground at Maidla during the 10th– 13th centuries. Newborn life expectancy was 25.7 years, the crude death rate was 38.9‰, and women gave birth to 4 to 5 children on average. This level of newborn life expectancy implies low child mortality and favourable living conditions, or the

47 under-representation of sub-adult (older than 5 years) burials in the grave. There are many possible reasons for child under-enumeration in the graves: the segregation in sub-adult burial practice is plausible (Allmäe, 2010), the crushing of cremains before burial or excavation technique. For example the crushing of burnt bones before burial in Late Iron Age western Estonia is probable (Allmäe, 2013); in this case, the fragile cremains of children become invisible in graves (Sigvallius, 1994: 32; Holck, 1997). In adequate excavation techniques are less plausible, as in the first Maidla grave the number of sub-adults is representative.

Table 5. Demographic data of Estonian graves

Rõsna- Rõsna- Suure- Name of the grave Maidla I Maidla II Sand Saare I Saare II Rõsna barrows Type of the grave stone-grave stone-grave sand barrow sand barrow sand barrow summarised Archaeochronological dates 5th–6th cc. 10th–13thcc. 6th– 8thcc. 7th–8th cc. 6th– 8th cc. 6th– 8th cc. 341–413 990–1120 420–540 390–540 430–580 390–580 AMS dates 68.2 % (range) AD AD AD AD AD AD 258–427 980–1160 390–570 340–540 420–600 340–600 AMS dates 95.4 % (range) AD AD AD AD AD AD Number of burials 20 74 61 46 40 147 Sex determination Males 5 13 14 10 8 32 Females 5 16 15 8 8 31 Index of masculinisation (%) 1.0 0.81 0.87 1.25 1 1.03 Undetermined sex 2 23 11 10 9 30 Age structure Subadults 0-15 (yrs) 8 22 21 18 15 54 0 0 7 5 5 3 13 1 – 6 3 6 6 6 4 16 5 – 9 4 8 7 3 5 15 10 – 14 1 1 3 4 3 10 Adults (yrs) 12 52 40 28 25 93 15 – 25 0 1 5 5 1 11 20 – 39 5 25 13 6 5 24 35 – 64 3 12 11 5 3 19 60+ 0 2 0 2 4 6 Adult 20 + 4 12 11 10 12 33 Proportin of subadults 40% 29.70% 34.40% 39.10% 37.50% 36.73 Proportion of adults 60% 70.30% 65.60% 60.90% 62.50% 63.27 T = usage of grave (yrs) 75 250* 150 150 150 150

48 Life tables, raw data after Acsady & Nemeskeri (1970).

0 e 0 newborn life expectancy 30.3 28.16 27.25 26.54 32.08 28.35

Crude death=crude birth rate 0.033 0.036 0.037 0.038 0.031 0.035 Population size ** 8.1 8.3 11.1 8.1 8.6 27.8 Corrected life tables after Rösing and Jankauskas (1997)

0 e 0 newborn life expectancy 19.92 19.15 18.67 19.44 22.01 19.75

Crude death=crude birth rate 0.050 0.052 0.054 0.051 0.045 0.051 Population size ** 8.2 8.5 11.3 8.3 8.8 28.3 After Boquet & Masset 1982 Juvenility index 5-14/20+ 0.43 0.18 0.28 0.31 0.31 0.30

0 e 0 newborn life expectancy 10.9 25.7 18.1 16.9 16.7 17.1

Crude death=crude birth rate 0.0921 0.0389 0.0551 0.0593 0.0600 0.0585 Population size ** 2.9 7.6 7.4 5.2 4.4 16.7 GRR - 2.5 3.5 3.5 4.1 3.8 TFR - 5.1 6.9 7.2 8.4 7.8 Population size ** 2.9–8.2 7.6-8.5 7.4-11.3 5.2-8.3 4.4-8.8 16.7-28.3 Reproduction after Henneberg (1976) Rpot 0.648784 0.5208709 0.5139019 0.543373 0.6789304 0.5671374 R0 (Uc=7.45) 1.45 1.18 1.25 1.23 1.58 1.34 C - average number of births 4.8 3.88 3.82 4.05 5.06 4.22 *after Mandel 2003 ** after Ubelaker 1989 Concerning community size, the results of osteological analysis of cremains from the Middle Iron Age barrow cemeteries at Rõsna (south-eastern Estonia) indicate the same pattern – one family or household used one barrow cemetery. The assumptions on community size based on the archeologically determined number of burials (Rõsna-Saare I and Rõsna-Saare II) and a mortality rate of 40‰ show similar results (Ligi, 1989; Lang, Ligi, 1991: 227); barrow cemeteries should be regarded as the burial places of a single family, or at least a small group of people (Lang, Ligi, 1991: 226–228; Ligi, 1989). Similar results have been obtained from Iron Age east Lithuania: community sizes of 5–15 individuals usually buried their dead in one barrow cemetery, this number of individuals corresponding to a group of people the size of the average nuclear family over several generations (Kurila, 2009). Newborn life expectancy at Rõsna during the Middle Iron Age was 17.1 years, crude death rate was 58.5‰ on average and women gave birth to 4.2–7.8 children, depending on the model, (Table 5), indicating relatively unfavourable living conditions in comparison with Late Iron Age Maidla.

49 The present study demonstrates that demographic figures depend on the model or sample chosen. The overall variability in the proportion of adults and sub-adults in graves is not striking, but when skeletal samples are very small, any minor change in proportions of age cohort has a significant impact on the demographic figures.

3.3. TEMPORAL AND CULTURAL DIFFERENCES IN MIDDLE AND LATE IRON AGE WEST ESTONIAN AND SOUTH-EASTERN ESTONIAN BURIAL CUSTOMS

3.3.1. The grave types, dates and burial practice The studied Iron Age graves in west and south-eastern Estonia are of different sizes, construction types and periods, and burial practice (cremation, inhumation) has also influenced the character and composition of bone material. Some of graves have been more or less thoroughly investigated (the graves at Maidla and Rõsna; and those at Kirbla and the Uugla stone graves respectively), while others (Põlgaste tarand grave, the flat-ground cemetery at Siksälä, and several stone-graves at Läänemaa) have been investigated only partially. Most of the Iron Age graves investigated during the present study have a construction that is well known in Estonia through the irregular stone graves at Läänemaa (west Estonia), the rounded and long sand barrows characteristic to the Long-Barrow Culture in south-eastern Estonia, the tarand graves in various forms characteristic of Estonia since the Early Iron Age in different parts of Estonia (Mandel, 2003; Lang, 2007a: 170, 203; Laul, 2001: 190–195; Aun, 1992, 2005). The Late Bronze Age stone-cist grave at Kaseküla (Mandel, 1975; Lang, 1996: 297) has been studied as a case of a special burial place for Late Iron Age infants (Allmäe, 2010). The flat-ground (or under level ground) cemeteries with cremations were not recovered since 2009–2010 at Siksälä in south-eastern Estonia and at Uugla, west Estonia (Valk, Allmäe, 2009, 2010; Mandel, 2011); the scattered cremains under level ground is a new type of burial ground for Estonia. In many west Estonian stone graves (Maidla I and II, Ehmja, Lihula) cremations as well as inhumations were practiced while at Uugla, Keskvere and Kirbla only cremations were present (Allmäe, 2003, 2013, 2014a; Mandel, 2003; Table 3). In the studied south-eastern Estonian graves cremation was prevalent (Aun, 1992; Laul, 2001; Allmäe, 2013, 2014b), except for the Kirikumägi flat-ground cemetery at Siksälä, where unburnt fragments of the skeleton of a 3- to 4-year old child were found (Valk, Allmäe, 2009, 2010). In Suure-Rõsna barrow cemetery, some unburnt human vertebrae were found in one bone assemblage, this could have happened due to unsuccessful cremation in which the temperature and oxygen flow were insufficient to properly burn the corpse. According to archaeochronological dates the tarand grave at Põlgaste is from the 3rd–5th cc. (Laul, 2001: 42), the Maidla I stone grave is from the 5th–6th cc. (Mandel,

50 2003) and the sand-barrow cemeteries at Rõsna, south-eastern Estonia, from the 6th–8th cc. (Aun, 1992). However, radiocarbon dates (AMS) of cremated bones added new information regarding the dating of the burial sites. Two samples of cremated bone dated cremation burials in the tarand grave at Põlgaste to the 1st century BC–2nd century AD; one sample dated cremation burial in the first stone grave at Maidla to the (3rd)4th–5th centuries AD, and four samples of cremated bone dated burials in the sand barrow cemeteries at Rõsna to the 4th–6th centuries AD. However, in comparison with archaeological chronology these burials sites could have been established somewhat earlier (Allmäe, 2013, 2014a). The usage periods of Maidla I stone grave and Rõsna sand barrows overlap despite the overall fragmentation of cremated bones and proportion of sub-adult burials are similar. The proportion of sub-adults was only 17.4% at Põlgaste tarand grave, although nothing certain can be concluded on the number of sub-adult burials, as the no all of the material was available for study. The cremated bones themselves from the Põlgaste tarand grave demonstrated peculiar characteristics. They exhibited very even colouration and non-typical fracture patterns that could be attributed to the cremation of the bones (Buikstra, Swegle, 1989; Whyte, 2001). Cremating dry bones is a phenomenon found earlier in connection with an Early Iron Age tarand grave at Tandemägi, Estonia (Kalman, 2000b; Lang, 2007a: 180), and, for example, detected in Early Bronze Age graves in Sassary, Italy (Piga et al, 2008). The custom of cremating human remains after the corpses had decayed is probably uncommon for the Middle Iron Age, although there is some evidence for this from Early Iron Age Estonia. However, Anu Kivirüüt (2014) has lately re-visited the osteological material from Tandemägi and disagrees with Kalman (2000b). Kivirüüt argues that the corpses have been burnt, but at low temperatures. Further studies of osteological materials from tarand graves are essential to unravel the variability in Early Iron Age burial custom. All other studied graves in west and south-eastern Estonia indicated colouration and fracture patterns characteristic of burning corpses.

3.3.2. The influence of the funeral pyre on human remains The effects of a funeral pyre on human bodies have been studied thoroughly for decades, thus the effects of temperature on bodies and bones are well known. The thermal alteration of bones due to temperature, oxygen flow and duration of burning are expressed in shrinkage, different fracture patterns and colouration of bones (Shipman et al, 1984; McKinley, 1989, 1994a, 1994b; Buikstra, Swegle, 1989; McCutcheon, 1992; Stiner et al, 1995; Warren, Maples, 1997; Bennet, 1997; Whyte, 2001; Wahl, 2008; Walker, Miller, 2005; Holck, 1997/2008: Walker et al, 2008, Symes et al, 2008; etc.). The observations and experiments in modern crematoria have demonstrated that cremation of an adult individual produces 3075 g/7.8 l of cremains on average, with the volume of cremains diminishing to 2.95 l after the bones are intentionally crushed. The more modern incinerators diminish the amount of cremains by one third, to approximately 2000–2500 g, giving good good proof that cremation is an extremely important factor (Holck, 1997/2008). Surprising variation in the weight

51 of cremains was found by Warren and Maples (1997). They found that the weight of cremains of fully developed adults ranged from 876 g to 3784 g, the average being 2430 g. Hence, we know that the heat of the funeral pyre results in fragmentation and remarkable variation in weight and volume of cremains. On one hand, these amounts of cremains are not often found in archaeological cremation burials; on the other hand, the bone fragment size of modern cremations is comparable to the ancient cremations buried in sealed containers (McKinley, 1994a). Having examined the heat of the funeral pyre, it would be safe to assume that the temperature of the funeral pyre is just one of the main reasons for the fragmentary nature of bones recovered from prehistoric graves and that cremains are subjected to factors other than heat.

3.3.3. Other agents responsible for bone fragmentation in recovered cremation burial deposits The fragmentation of cremated bones after the funeral pyre is caused by cremain handling, i.e. cooling, raking, collecting, sorting, transporting and burying. Ritual crushing of cremated bones before burial has also been suggested. Further fragmentation can be the result of the excavation of the material, as well as subsequent sieving, cleaning and washing during post-excavation processes, as well as when depositing and analysing the bone material (Lange et al, 1987; McKinley, 1989, 1994a, 1994b; Formisto, 1996: 87; Sigvallius, 1994; Kaliff, 1997: 70; Kaliff, Oestigaard, 2004; Holck, 1997/2008: 35; Harvig et al, 2012, 2014a; Harvig, Lynnerup, 2013; Pankowská et al, 2014; Harvig, 2015). It has been argued that taphonomic processes in soil and both the excavation and post-excavation activities can cause the largest amount of damage to archaeologically recovered cremains rather than coming from ritual behaviour in the past (McKinley, 1994a; Formisto, 1996; Harvig, Lynnerup, 2013; Harvig, 2015). The influence of archaeological excavation on the degree of cremain fragmentation depends on the soil type, excavation methodology, time of excavation, and other possible factors (Harvig et al, 2012). The maximum size of the cremated bone fragment recovered from modern crematoria is 25 cm, immediately after raking cremains the maximum length reduces (McKinley, 1994c), which means that all activities after the funeral pyre and before burial cause extra fragmentation. Further fragmentation depends on when and how the bones are handled. The hot bones are brittle, cool ones are relatively hardy (Harvig, 2015). When dealing with archaeologically recovered cremated bones it should be noted at this point that bone fragment size means here the size after excavation rather than the size of the deposited bones (McKinley, 1994c). This has been described well in studies by Danish researchers on how micro-excavation of burial urn affects bone fragment size (Harvig et al, 2014a). Later on, several possible agents appear that affect cremated human bones before they are analysed by an osteoarchaeologist. Pankowská and colleagues (2014) have found laboratory techniques to be the most destructive; for example in sieved cremains the bone fragments are approximately

52 three times smaller than materials from micro-excavation. Therefore, the idea of the ritual crushing of bones has been neglected by many authors, with good reason. Finnish scholars (Wickholm, Raninen, 2006) have pointed out that even if the ritual activities are difficult to prove, they should not be excluded from burial archaeology, despite activities after the funeral pyre and before burial being in a grey area for ancient cremations. The problems of research into ancient burial customs and rituals are probably the reason why contextual taphonomy is gaining popularity. The main purpose of contextual taphonomy is to distinguish cultural, post-depositional, excavation and post-excavation factors (Pankowská et al, 2014). We can hypothesise that apart from taphonomic processes in soil, excavation and post-excavation activities, the fragmentation stage of cremains, the amount discovered in-grave reveals some other factors. For example 1) the burial custom itself - scattered cremations, compact nested burial of cremains (the latter sometimes in some kind of container), burial of cremains in several places; 2) the variability of actions with cremains before burial – collection, sorting, division and manipulation of bones. Can we prove ritual activity with the help osteological research? We do not know how our ancestors thought and acted during the death and funeral of a family or community member. We cannot perceive their beliefs and we will never have a clear vision of their burial customs. Therefore partial burials of cremated bones or the burial of bones in several places is intriguing, although not easily detectable using archaeological or osteological methods. The basic variables of cremation deposits are the representation of cremains, their volume and weight, and the degree of fragmentation (McKinley, 1989). The degree of fragmentation can be measured directly – for example measuring the size of bone fragments (Lahtiperä, 1970; Iregren, 1972; McKinley, 1994a, 1994b, 1994c; Formisto, 1996; Allmäe, 2003, 2013; Allmäe, Maldre, 2005, Valk, Allmäe, 2010) or relatively, using the ratio of weight and volume of cremains (Harvig, Lynnerup, 2013; Pankowská et al, 2014). The latter is not applicable to scattered human remains, or in other cases, where burials are not discernible. Unfortunately the Estonian material was not suitable for weighing and volumetric measurement because of the collective nature of burials with scattered cremains and the mixed nature of animal and human bones in barrow cemeteries, where the all fragments cannot be sorted according to species. For Estonian material the measurements of long bone and cranial fragments was conducted to estimate the fragmentation stage. It should be noted that Estonian materials are not usually sieved during excavation and neither are they washed and sieved afterwards. This means that the post- excavation effects are not comparable to those caused by the method used in, for example, Britain (McKinley, 1994c). In Estonia the fragmentation is less pronounced in graves from the Early to Middle Iron Age, despite the grave construction - sand barrow, tarand grave or irregular

53 stone grave. The graves with higher fragmentation of cremains are mainly from the Late Iron Age regardless of whether the burnt bones were recovered from sandy soil, clay-rich heavy soil or from stone-graves. The crushing of bones has been the subject of discussion for decades. Some authors argue that it is part of a burial ritual (Holck, 1997/2008: 35; Kaliff, 1992, 1997; Sigvallius, 1994; Kaliff, Oestigaard, 2004), while others rationalise the subject and refer on effect of taphonomic processes in situ, or propose that post-excavation activities are responsible for heavily fragmented bones (McKinley, 1994a; Formisto, 1996, Harvig et al, 2012; Harvig, Lynnerup, 2013, Pankowská et al, 2014, Harvig, 2015). Based on research into Danish Late Bronze Age urn graves Harvig and colleagues (2012) have rejected the idea that cremains were deliberately crushed before burial in an urn or that some kind of deliberate insertion of different bones was carried out during the funeral. We could argue the same for south-eastern sand barrow cemeteries; here the crushing of cremated bones was not practiced as evidenced by the fact that the bone units often contain large spongy parts of bones, which are easy to crush even today. Later on Harvig and colleagues (2013, 2014a) compared cremated bone materials from Danish Bronze Age urn burials and Early Iron Age cremation pits. The former contains well preserved cremains of individuals; the latter demonstrates a lack of some anatomical areas, extreme fragmentation and heavy wear. The authors reached the conclusion that the extreme fragmentation of cremains in Danish Early Iron Age cremation pits points to handling of cremains before deposition and is not caused only by post-depositional taphonomic processes; they defined cremation pits as deliberate sorted deposits, because usually no pyre debris is found amongst the bones (Harvig et al, 2014a). Proving ritual crushing of bones in west Estonian Late Iron Age stone graves with scattered cremains is complicated, because it's difficult to distinguish taphonomic wear from the results of ritual crushing; the same has been concluded for Danish Early Iron pit graves (Harvig, 2015). According to McKinley and Bond (McKinley 1994b, 1994c; McKinley, Bond, 2001) the fragmentation of bone occurs for several reasons, for example from the raking of the burning remains during the cremation process, of the later collection and subsequent interment of remains, making it difficult to assess whether bone was deliberately fragmented as part of the cremation ritual. Therefore we cannot rule out the ritual crushing of cremains in Late Iron Age west Estonia, although neither can we confirm it.

3.3.4. The character of cremated bones and burial customs in Iron Age Estonia Tarand graves were used in Estonia from the Pre-Roman Iron Age (Lang, 2007a: 170, 188, 2007b: 104, 122), becoming the most widespread grave type during the Roman Iron Age (Lang, 2007a: 219, 2007b: 126). The osteological material from north Estonian tarand graves have been well studied over recent decades (Kalling, 1993; Kalman 2000a, 2000b, 2000c, 2000d; Kivirüüt, 2011, 2014), although very

54 little is known about the cremated bones in south-eastern Estonian tarand graves. Amongst the south-eastern Estonian cremation cemeteries studied, the tarand grave at Põlgaste (radiocarbon dated from the 1st century BC to the beginning of the 2nd century AD), demonstrated osteological features not very well recognised in Estonian archaeology. The cremated bones exhibited quite large bone fragments and, while the majority of bone units were determinable, the number of estimated remains of sub-adults was modest (17.4%). This would seem to suggest that sub- adults were not commonly buried in that tarand grave, or that the part of the grave that consisted of sub-adults was not available for osteological analysis. The colouration of the cremains was evenly pale with a light brown and occasional orange-red hue and the fracture pattern was not typical of cremated corpses. It should also be noted that only some of the bone fragments had been calcinated. The secondary burial custom is suggested because of osteological analysis. The bodies should have decayed to some extent before the cremation of the remains and therefore before burial in the tarand grave. The burial custom according to which fragments are carefully picked from pyre site and buried as compact nested bone assemblages is indicated by less fragmentation, better preservation of the spongy parts of bones, a greater number of determined bone units, and a higher proportion of identified individuals, including the cremated remains of infants and children (Allmäe, 2013). The latter is characteristic of Estonian (4th)6th–8th century sand barrow cemeteries with cremations. Mare Aun, long-standing researcher into the Long Barrow Culture, has observed various burial practices in sand barrow cemeteries. The cremated bones are placed into barrows and sometimes also buried under the barrows in clay, wooden or other organic containers, or simply laid in as assemblages of cremains (Aun, 2005: 106, 2006: 115). It should also be noted that pyre sites have not been found near barrow cemeteries – the cremations themselves took place somewhere else (Aun, 2006: 177), and the cremains were collected and carried to the burial place. In barrow cemeteries human remains pointed to two different contemporaneous burial customs. The first was of careful collection of bones from the pyre site, bone units that are of white or yellow colour and do not contain soot, rather these assemblages consist of clean bones. The other custom was to collect and bury cremated bones with pyre debris, soot and charcoal, in which case ‘sooty’ bone assemblages were detected in graves. In west Estonia the first Maidla stone grave from the (4th)5th–6th centuries that has scattered human cremains is contemporaneous with sand barrow cemeteries in south-eastern Estonia. The common characteristics of these Middle Iron Age graves are relatively low fragmentation of cremated bones and a comparable proportion of sub-adult burials (Allmäe, 2013). We may speculate that the similarity could result from the simultaneity of the graves and a comparable community attitude towards children. The third burial custom observed was the scattering of the cremains of several people over the grave. This custom causes the dispersal of bone fragments in a way

55 that it produces high numbers of small non-determined bone units, making the determination of the number of individuals in the grave difficult. The phenomenon to scatter human cremains all over the burial place is deliberate and attributed to the burial custom of collective nature (Purhonen, 1996: 120; Wickholm, Raninen, 2006; Mägi, 2007, 2013). The goal was to mix the remains in the way that individual graves are not discernible. The scattering of the human remains and all other pyre goods after the funeral pyre had burnt down has been attributed to ritualistic activity the purpose of which was to emphasise the collective nature of the ancestors (Wessmann, 2010: 58). Obviously the bones have been manipulated after cremation in west Estonia. Unfortunately the possible way in which cremains were influenced by the stone structure of the grave was not observed during excavation. Therefore we do not know if the bones were scattered between existing stone constructions or if the remains were later covered by the stones. However, characteristics of graves with scattered cremations were observed at the second Maidla stone grave: bigger units of bone were found outside the pyre sites and most artefact complexes were not related to bone assemblages (Allmäe, 2003: 247). Only one correlation was found in the graves in the form of spearheads related to the estimated concentration areas of cremated human bones (Mandel, 2003: 135; Allmäe, 2004). Intrasite spatial analysis of the 10th–13th century cemetery at Madi (south Estonia), which has dispersed cremations, has also shown no clustering of the artefacts and burnt bones in-grave (Konsa, 2013), and neither do the distribution patterns of artefacts and burnt bones overlap in most cases here. Konsa has suggested that the cremated bones have been distinguished from the rest of the pyre remains and treated separately from grave goods (Konsa, 2013). As we see, graves with scattered human cremains were not only found in west Estonia, but also in south and south-eastern Estonia. In south-eastern Estonia, at Kirikumägi, Siksälä archaeologist Heiki Valk discovered a flat-ground cemetery from the 10th–12th centuries with scattered human cremains. Here the cremation grade and the fragmentation stage of cremated bones is similar to Late Iron Age stone graves in west Estonia (Valk, Allmäe, 2009, 2010), despite the different grave construction (stone grave in west Estonia, flat-ground cemetery without stone constructions at Siksälä) and different soil conditions. In west Estonia the soil is thick and clay-rich, in addition to which later excavation activities could easily have destroyed the spongy parts of bones. At Siksälä the soil is light and sandy and excavation hardly damaged the bones. Despite the differences in grave construction and soil condition the size of bone fragment at Siksälä flat-ground cemetery is comparable to those in Late Iron Age west Estonian stone graves. The similarity in bone fragmentation could be due to the handling of bones before burial, for example crushing. The pyre site itself was not found in the grave area at Siksälä, although this could be due to the fact that the excavation area is quite small and the cremation place itself has not been found yet, or it may be because the pyre site is not in the grave area as we see in the case of sand barrow cemeteries at Rõsna.

56 Apart from Estonia cremation cemeteries under level ground (cemeteries without formal structure) with scattered human remains are also characteristic to Karelia, Curonia and Finland (Mägi, 2002: 24, 2013: 185; Wickholm, 2008; Wessmann, 2010: 19). In recent decades the results of osteological analysis of cremated bones have been published only for two Finnish graves. The first is cremation cemetery A at Vainionmäki, Laitila (600–800 AD), excavated between 1986 and 1994. Formisto (1996: 81–87) has suggested that the cremation itself was conducted in the grave area. For several west Estonian cremation graves Mandel (2003) has suggested the same. Mandel (2003) has described heavily burned areas that have a greasy charcoal layer in the grave area. This is in accordance with criteria previously established by other researches (Iregren, 1972: 73; McKinley, 1989; Arcini, 2005) to determine the areas where corpses were burned in the grave area. The average size of the longest bone fragments measured in the Vainionmäki A grave was 1–3 cm, with a maximum length of about 8 cm. Bone colour is described as grey, although the bones were covered with soot, which sometimes caused difficulties in determining the actual colour (Formisto, 1996: 81–87). Despite the time gap, the Vainionmäki A cremation grave, with irregular stone settings, scattered cremains, many sooty bones and comparable bone fragmentation stage is very similar to west Estonian graves, especially the Maidla II stone grave from the 10th–13th centuries. The second Finnish example is Rikala, under the level-ground cremation cemetery at Salo (initially dated to the Merovingian period). The osteological study and radiocarbon dating of cremated bones from Rikala (section VII) revealed that this grave had quite a complicated nature, for example the scattered human remains and multiperiod use of the grave. According to new research the grave was used from the second half of the 1st century to the beginning of 7th century AD (Mäntyla- Asplund, Storå, 2010). Despite the time gap there are some similarities between Rikala at Salo (1st–7th cc.) and west Estonian Late Iron Age stone graves with irregular stone settings, or with the flat-ground cemetery Kirikumägi at Siksälä (10th–12th cc., south-eastern Estonia). These are cremation cemeteries under level ground (flat-ground cemeteries). Kirikumägi and Rikala (Mäntyla-Asplund, Storå, 2010; Valk, Allmäe, 2010) show no trace of funeral pyres in the grave area. Partial excavation of these two graves could also be the reason why pyre sites are not detected at Rikala, Salo, or at Kirikumägi at Siksälä. However, it is possible that the funeral pyre was arranged far away from the grave area in both cases. Further large scale studies could reveal the peculiarities of mortuary practices relating to under level ground cemeteries. The cremated bone material from Late Iron Age graves at Läänemaa (west Estonia) and Siksälä (south-eastern Estonia) are more fragmented, indicating a smaller proportion of bone fragments. The bones themselves are often sooty, in many cases the pyre sites are found in the grave area with, occasionally, the cremated remains of infants and children found. We may assume that the character of the cremated bones is caused by the mortuary practise of the Late Iron Age, i.e. scattering

57 cremains with pyre debris into the grave area, although we cannot rule out the possibility that bones were crushed before burial, or that some bone parts were taken from the pyre site and then buried somewhere else. Graves with few human bones, or a total absence of them, have already been described, often along with reasoning about distribution of cremains between several burial and/or ritual places (Gejvall, 1961; McKinley, 1989; Kaliff, 1992: 121–122; Sigvallius, 1994: 27–32; Parker Pearson, 1999; Allmäe, 2003: 258; Wickholm, Raninen, 2003: 4; Kaliff, Oestigaard, 2004; Williams, 2008; Wickholm, 2008: 90; Wessmann, 2010: 53–54; Tvauri, 2012: 194; Veselka, Lemmers, 2014). ‘Grave’ as a concept can be symbolic, although it can be difficult to distinguish the pyre site from the burial site or cult place.

3.3.5. Pyre sites Caroline Arcini (2005) has argued that osteological materials recovered from pyre sites are often characterised by a lack of the petrous parts of the temporal bone. In light of this observation we can speculate that the Late Iron Age graves at Kirbla and Uugla (I, II, III), which have burned areas with greasy black soil, commonly with small pieces of artefact and heavily cremated bones, are mainly excavated pyre sites with a surrounding area of scattered cremains. The excavations of stone graves in Läänemaa have revealed several pyre sites inside grave areas (Mandel, 2003: 39– 47, 70–73, 75–80, 82, 90; Mandel, Allmäe, 2009). Analysis of osteological material showed that petrous parts of the temporal bones in the Uugla graves were mainly found in pieces, while not even this was found in the Kirbla grave (Allmäe, 2003: 246–247; Mandel, Allmäe, 2009). Uugla II is stone-circle grave, the only example found in west Estonia so far. The cremation was carried out inside a stone circle of 2 m x 1.6 m on site (Mandel, 2003: 82). At Kirbla cremated human bones were mainly found in connection with a sooty patch of 3 m x 2.2 m (Mandel 2003, 90). At Uugla I and III very fragmentary bone material was also mainly found in sooty areas (Mandel, 2003: 82; Allmäe, 2003: 247; Mandel, Allmäe, 2009). Where was the bulk of the human remains found at the Uugla pyre site? Mati Mandel (2011) has found cremated bones under level ground around the Uugla III grave; here cremains were found associated with sooty soil patches or in small pits, no artefacts were found with bones. All three small Uugla (I–III) stone graves are all very close to each other, although we could suggest that Uugla is similar to the Maidla II grave field, where human remains were cremated in the grave area and then sorted and scattered or buried near the pyre sites. The second important observation made by Arcini (2005) is that pyre sites may contain a relatively large number of animal bones or even whole skeletons, while at the same time very few human bones are detected. These features are in some extent characteristic of Estonian Viking Age stone-circle graves at Piila, Saaremaa Island (Mägi et al, 1998), especially grave IV: 2, where a number of dog bones where recovered but only a few human bones. Cremated dogs have been found in association with sooty patches in Late Iron Age stone graves at Läänemaa, for

58 example at Uugla I and III and Kirbla. Cremated dogs have also been found in the same circumstances at Maidla II (Maldre, 2003: 269–274). We might hypothesise that at Piila human cremains were carefully picked from the pyre site and buried elsewhere or otherwise, and that some cremated human bones with artefacts were brought to cult or burial place from pyre sites, as suggested by Marika Mägi (2007). It should be noted that Mägi et al (1998) found the probable pyre site for mainly human remains in grave IV: 4 at Piila, where a 10 cm thick sooty layer with cremated bones was found (Mägi et al, 1998). Arcini (2005) has pointed out that “the belief that pyre sites were collective and used repeatedly has led to the search for larger pyre sites with traces of repeated cremations, which is expected to generate an abundance of charcoal and fire-burnt soil.” However, this is not always the case because cremation conducted at one site only once and left uncovered leaves no remarkable traces. The remarkably sooty and greasy patches of soil, or even areas at the Läänemaa stone graves are obviously pyre sites used recurrently. Pyre sites were not found in the grave area in south-eastern Estonia and it is possible that they were far away from the cemetery, for example near settlements, where it was easier to control the funeral pyre during the cremation process. It is not unusual that pyre sites are not found in the grave area; even at Spong Hill, the largest excavated Anglo-Saxon cremation cemetery in the UK, no pyre sites were found (McKinley 1994c). On the other hand, most Romano-British cemeteries probably functioned as crematoria, although the evidence of actual pyre sites is scarce (McKinley, 2008). Cremation in the grave area has also been suggested for Vainionmäki A cremation cemetery (Formisto, 1996: 81–87). The cremation technology, pyre sites and their relationship to burial custom in prehistoric Estonia is a subject worthy of further study, certainly to date we have very little knowledge of this subject.

3.3.6. Cemetery and community size The size of community using one burial place has been an important question for social archaeology. Previous demographic studies of Estonian prehistoric and historic periods have been based on different assumptions and modelled using various methods. The first demographic model used to calculate community sizes for prehistoric Estonian communities were based on the number of nested bone assemblages in sand barrows (Ligi, 1989) and the number of artefacts in tarand graves, using a presumed mortality rate 40‰. The estimated average size of community that used one tarand grave was 5–9 individuals in north-eastern, south- eastern and central Estonia, and 3–4 individuals in north-western Estonia (Lang, Ligi, 1991: 224–225). Valter Lang (1996: 375) subsequently corrected these numbers to 7–13 and 3–8 individuals respectively. A community size of 8–10 individuals was obtained from skeletal remains from Roman Iron Age tarand grave I at Viimsi in northern Estonia (Lang, 1993: 56), which is in good accordance with artefact-based calculations.

59 Osteological research into skeletal materials from Pre-Roman Iron Age tarand graves at Poanse (west Estonia) suggested a community of 4–6 individuals depending on the period of grave use (Kalman 2000a; Lang 2007a, 224), while the community at Tandemägi (Võhma, Northern Estonia) consisted of 6 individuals (Kalman, 2000b; Lang, 2000: 206). Lang (2007a: 224–225, 2007b: 357) pointed out that one family used the tarand graves for centuries and that the custom of burying only members of the nuclear family, and sometimes only some of them, in stone graves was practised during the Pre-Roman and Roman Iron Ages in Estonia. The demographic models used to assess the size of living population based on Middle Iron Age ((4th)6th–8th cc.) cemeteries at Rõsna and at Maidla demonstrate that cemeteries were commonly used by one family or household (Allmäe, 2014a, 2014b). Similar results have been obtained in Sweden, Norway and Lithuania (Iregren, 1972: 59–62, 120; Holck, 1997/2008: 59; Kurila, 2009). Based on the analyses of osteological materials the Late Iron Age stone-grave at Maidla, dating from the 10th–13th centuries, was also used by one family or household (Allmäe, 2006, 2014a). According to archaeological data the same was found for Late Iron Age stone-graves on Saaremaa, where generally one, but sometimes 2 or 3 families, used one cemetery (Mägi, 2002: 74). Hence, the demographic calculations based on human skeletal remains tend to support earlier demographic assumptions based on the archaeological and osteoarchaeoalogical materials under study. The present study also demonstrates that demographic figures depend on the model or sample chosen. The overall variability in the proportion of adults and sub-adults in graves is not striking, but when skeletal samples are very small, any minor change in proportions of age cohorts has a significant impact on demographic figures.

3.3.7. Burial customs for children and infants There is a common argument that in some burial places the number of infants and children is under-represented, both for inhumation as well as cremation burials. The proposed reasons for this vary from suggesting that the bones of infants are too fragile to survive in unsuitable soil conditions (Walker et al, 1988; Guy et al, 1997; Buckberry, 2000), that infants may have been buried elsewhere or have been lost from the cemeteries because of their continuous use, or that excavation techniques were responsible (Molleson, 1991; Saunders, 1992; Roberts, Manchester, 1995; Mays, 2006; Chamberlain, 2006). The disappearance of infant and child remains from cremation graves is more likely to be due to the destructive effects of pyre temperature, the extreme fragility of burned infant bones and later taphonomic factors in the grave (Holck, 1997; Waterhouse, 2013; Jaeger, Johansen, 2013). Very few sub-adult burials have been found amongst Iron Age cremation burials (Iregren, 1972: 50; Holck, 1997/2008: 63, 119; Sigvallius, 1994: 40).

60 Osteological analysis of 5th–13th century west Estonian graves demonstrates that 28.1% of burials belong to children, 51.9% of inhumations, and only 12.3% of cremations (Allmäe, 2003). In south-eastern Estonian sand barrows of the (4th)6th– 8th centuries, the proportion of sub-adult cremations varies between 34.4–39.1% (Allmäe, 2014a, 2014b; Table 5). Infant cremations were totally absent in west Estonia, but were relatively numerous in south-eastern barrow cemeteries (Allmäe, Maldre, 2005; Allmäe et al, 2008), raising the question: is this a regional or temporal difference in the burial practice of sub-adults? The Long Barrow Culture sand-barrows that have been investigated are from the (4th)6th–8th centuries, although the west Estonian graves are usually from Late Iron Age. The only exception is the first Maidla stone grave, also dated to the (4th)5th–6th centuries. Here the proportion of child cremations is quite representative – six burials from a total of 20. For example, in the second Maidla grave, from Late Iron Age, the number of detected child cremations is only two (74 burials in total). With this surprisingly small number of child cremations in mind, the Kaseküla stone-cist grave, with its remarkable number of buried infants (Kalman, 2000d), caught the author’s attention because next to the Bronze Age stone-cist there is a Kaseküla 14th century stone-grave from Late Iron Age that has cremations (Mandel, 1975: 76, 2003: 108). In connection with the missing child burials in Late Iron Age west Estonian stone-graves the possibility to date the adult and infant inhumations in the stone-cist grave was intriguing: are the remains of infants buried at the same time with the adult in the cist? The results of radiocarbon dating and osteological analyses demonstrated that the infant remains in the Kaseküla stone-cist grave originate from the Late Iron Age, when the prevailing mortuary custom in West Estonia was cremation (Table 4; Allmäe, 2010). The later thorough radiocarbon dating of Kaseküla stone-cist grave confirmed the recurrent use of the burial place between the Bronze Age and Late Iron Age (Laneman, 2012). Secondary burials, or reuse of prehistoric stone graves, was known to Estonian archaeology, although these sites had not been radiocarbon dated or discussed comprehensively (Jaanits et al, 1982: 150, 178; Lang, 2000: 147, 2007a: 57–59, 191–192; Mandel, 2003: 140; Vedru, 2009). In relation to this, Tõnno Jonuks (2009: 175) has pointed out that the importance of once established burial places for our ancestors can be connected particularly to secondary burials in the graves. The same idea has been expressed by other researchers (Wickholm, Raninen, 2006; Wickholm, 2007, 2008; Wessman, 2010). Recently the radiocarbon dating of the Muuksi stone-cist and Jaani stone-grave at Väo (both in northern Estonia) have confirmed the recurrent use of one burial places from the Bronze Age to Late Iron Age in Estonia (Laneman, Lang, 2013; Laneman et al, 2015). Taking into account the possibility that infant and child remains were not cremated and not buried in family graves as a rule in west Estonia, but rather somewhere else, the Kaseküla stone-cist grave area might be an alternative burial place for community infants at Kaseküla during the Late Iron Age. The stone-cists graves are visible in the contemporary landscape as round heaps covered with sod that are

61 approximately 10–15 m in diameter and up to 1.5 m high (Lang, 2007a: 148). The significance of landscapes and monumental graves for our ancestors has been discussed by Gurly Vedru (2009), who suggested that the later reuse of burial sites was due to the people’s need to manifest themselves through ancestors and/or place. It is possible that the act of burial also served to physically ‘plant’ the dead into the land, making the remains an inalienable and fixed part of that land (Parker Pearson, 1999: 17). The grave at Kaseküla had probably acquired a different significance after 2000 years, but we do not know the reason why. Unfortunately we can only surmise what our ancestors thought at the time. Was the Kaseküla stone-cist grave a secret place to bury still-born and unexpectedly deceased infants, or to hide the consequences of infanticide? Ideas about why infants have been treated differently and were buried in unusual places in prehistoric times has been discussed by many authors. Proposals vary from human sacrifice to hiding the remains of unwanted children (victims of infanticide) to more practical reasons, like infant cremation being too time and resource intense (Holck, 1997; Green, 1998; Purhonen, 1999; Parker Pearson, 1999; Mays, 2006; Chamberlain, 2006; Faerman, Smith, 2008; Smith et al, 2011). Several reasons for infanticide have been suggested, including adaptive behaviour under certain conditions, such as ritual sacrifice, the need to control fertility and reproduction, control resources, eliminate disabled offspring, manipulate sex ratios and eliminate possible illegitimate offspring (Chamberlain, 2006: 171). However, we cannot know the beliefs, thought patterns or reasoning of our ancestors and so the question of why infants were buried in Bronze Age stone- cist is still unanswered many centuries later. The use of earlier stone graves is well-known in Sweden (Welinder, 1998: 188– 189), where a similarity to the Kaseküla grave can for example be observed in the Bronze Age cairns in Gotland. Here tens of infants were buried in the cairns, radiocarbon dating suggesting that they were used during the Middle Ages as an alternative to Christian churchyards (Lindqvist, 1981). The use of old graves to bury the remains of infants and small children is not an uncommon habit in human societies. Cillinis – special burial grounds for children – are well known from early Christian Ireland, for example, and the differential treatment of children’s remains in the Neolithic and Bronze Age is recorded through the reuse of earlier monumental graves, such as Megalithic tombs (Finlay, 2000). The use of Megaliths as burial grounds for infants, and for other ambiguous categories of individual, leads to a re-think of both the meaning and significance of the original monument (Finlay, 2000: 420). In Christian societies, un-baptised people, suicide victims and other people who violated Christian practices were not buried in consecrated land around churches and chapels (Purhonen, 1998: 119– 120). This could introduce a necessity for separate burial grounds and different funerary practices for these members of society. The radiocarbon dating of one infant inhumation from the second stone grave at Maidla also showed medieval dating (15th–16th cc. AD), hence the reuse of stone graves to bury un-baptised infants is not ruled out in west Estonian prehistoric graves (Allmäe, 2014a). The use

62 of old village cemeteries during the Christian period has probably been more common than archaeologists usually believe (Wickholm, 2007: 114); for example radiocarbon dating showed that Rikala cremation cemetery was used for much longer than previously thought (Mäntyla-Asplund, Storå, 2010). A remarkable number of infant burials has been found in the Lastekangrud Bronze Age stone-cist grave at Rebala (northern Estonia). In fact only young individuals were found buried here; moreover, 15 burials out of forty belong to infants (Kalman, 2000d; Lang et al, 2001: 39–45). The burials at Lastekangrud are not radiocarbon dated yet, although future research results from this grave could be illuminating. The skeletal evidence of children and juveniles from graves is crucial to understand the burial custom of children. Becker (2007) has emphasised that special attention should be paid to sub-adult skeletal material, as children's role in society change with age. Stig Welinder (1998: 187) has argued that when all modes of burial are taken into account there is no notable lack of children’s graves. Usually archaeologists focus on one grave type and archaeological period to find connections between contemporaneous burials in different grave forms, neglecting surrounding burials places. In the Culture of Long Barrows from (4th)6th–8th centuries in Northern Setumaa (south-eastern Estonia) the remains of deceased infants and children were commonly cremated and buried in family or community graves (Allmäe, Maldre, 2005, Allmäe et al, 2007a, 2008; Aun et al, 2008, Allmäe, 2014b). There are some barrows at Rõsna-Saare I and II cemetery where only the cremains of infants and children were buried, thus segregation by age exists within burial the ground here. For example in Rõsna-Saare I cemetery, in barrow number 2, a single child was buried, while in barrow number 9 five children were buried. In Rõsna-Saare II cemetery a single child less than 6 years old was buried in the 6th barrow (the longest one in the cemetery) in a wooden construction and with remarkable grave goods. In Rõsna-Saare II cemetery in the 4th barrow a new-born baby, and in the 11th barrow a 5 year old child were the buried with grave goods; no other burials were found here (Allmäe et al, 2007a). The relatively high percentage of child cremations in south-eastern sand barrow cemeteries (34.4–39.1%) shows something about local burial customs and sociocultural behaviour. The cremains of infants have been found in Long Barrow Culture sand barrows in Pskov, Russia (Khvoshchinskaya, 2004). The differences in the west and south-eastern Estonian burial practice of children and infants are obvious. Cremation burials in Maidla II began during the 10th century (Mandel, 2003: 59). Mandel has suggested (2003: 176) that the inhumations in grave 2 in Maidla could have been later ones, probably from the end of the 12th century or the first half of 13thcentury. The number of child cremations was low in Maidla second stone-grave from the Late Iron Age, whereas the number of child inhumations was rather representative (Allmäe, 2003, 2006, 2013, 2014a, 2014b). The radiocarbon dating of one infant inhumation from Maidla second grave (Allmäe, 2014a) showed the 15th–16th centuries, thus secondary infant burials in the

63 grave are likely, thus cremation and inhumation burials of children were probably not practiced in parallel. The occurrence of secondary infant burials in Maidla second grave demonstrated the same pattern seen at Kaseküla – later reuse of the grave, although of course the time-span between primary and secondary burials at Kaseküla is remarkably longer. The same phenomenon of reuse of under level- ground cemeteries is also known in Finland (Wickholm, 2008; Mäntyla-Asplund, Storå, 2010). After the cremation process only 1% of the body weight of foetuses and 2.5% of children are found (Warren, Maples, 1997). It’s highly likely, that the cremains of infants and children have mainly been lost due to in-grave taphonomic processes. If the bones were ritually crushed before burial then archaeologists are not able to find the cremation burials of infants and small children. It should be noted that heavily burnt white or pale bones from south-eastern sand barrows are very fragile: crushing the spongy parts of the bones is easy. Despite this the number of infant burials is quite representative in sand barrow cemeteries. In west Estonian Late Iron Age graves fewer spongy parts of bones are found, as well as fewer child cremains. If human remains, cremated or not, were buried somewhere else they have also been lost to further study. The difference in mortuary practice observable through the presence of infant and child cremations in graves is cultural as well as temporal. More precisely, the cremains of infants and children are often found in the graves of the Long Barrow Culture in south-eastern Estonia in the (4th)6th–8th centuries and are well represented in the (4th)5th–6th century stone grave at Maidla I. In Late Iron Age stone graves infant and child cremations are scarce, due to burial practice or taphonomic processes in the grave. The inhumation burials of infants and children in Late Iron Age graves are probably from later periods, from the end of the Late Iron Age and sometimes even from the Medieval period.

64 CONCLUSIONS

Fifteen Estonian graves with cremations and inhumations were investigated: ten graves (2008 bone units) from west and five graves (548 bone units) from south- eastern Estonia. During the research period various descriptive and metric data on cremated bone materials were collected. In addition, the results of biological age at death and sex of buried individuals, and estimates of the minimum and plausible number of burials enabled us to describe the composition of skeletal populations buried in different graves and to model some demographic figures for living communities. The collected data and research results were systematised and interpreted in order to provide some conclusions on Estonian cremations and burial customs during the Middle and Late Iron Age in west and south-eastern Estonia. The Accelerated Mass Spectrometry (AMS) method was used to radiocarbon date cremated and non-cremated bones from seven investigated graves. The radiocarbon dates helped to specify the time span over which the grave was in use and to find out if graves had been re-used. The AMS dates of cremated bones from the Maidla stone graves indicate a somewhat earlier establishment of the first Maidla grave in comparison with archaeological dating (5th–6th cc.). The radiocarbon date of cremated bone from Maidla second grave confirmed Late Iron Age (10th–12th cc.) cremation burials in the grave. One radiocarbon dated infant inhumation from the second Maidla stone grave shows that apart from the 12th–13th century inhumations, some of the infant inhumation(s) could be from the Medieval period. The infant inhumations in the Kaseküla stone-cist grave at Läänemaa (west Estonia) were radiocarbon dated to specify when these burials took place, and whether the infant burials were contemporaneous with Bronze Age burials or from later periods. The results of radiocarbon dating demonstrated that the infants were buried during the Late Iron Age, 2000 years after the adult in the stone-cist. In this way a distinctive mortuary custom was discovered for Late Iron Age infants in Läänemaa. The sand barrows from south-eastern Estonia are archaeologically dated to the 6th– 8th centuries, with radiocarbon dates of four cremated bone samples giving the 4th– 6th centuries. Two samples of cremated bone from the 3rd–5th cc. tarand grave at Põlgaste were dated to the second half of the 1st century BC to the beginning of the 2nd century AD. Radiocarbon dating indicated the possibility of a somewhat earlier establishment and use of the graves in comparison with archaeochronological dating in south-eastern Estonia. To establish a reliable chronology for the studied graves in the future the number of samples that are radiocarbon dated should be far more representative. A comparative study of ten graves on the basis of the dates, bone fragmentation and number of in-grave determined bone finds was conducted. The study showed that there is a difference in the proportion of determined bone units between the 10th– 13th cc. AD and the 1st c. BC – 8th c. AD graves, the latter indicating higher proportions. The graves from the 1st c. BC – 8th c. AD also indicated less bone fragmentation. Both measurements (cranial and long-bone fragments) reduce over

65 time – the earlier graves with burials of compact bone assemblages show a lower stage of fragmentation than the later ones. This also confirms that the custom of scattered cremation burials is distinguishable on the basis of higher bone fragmentation and lower success of identifications. Nearly all bone materials demonstrated fracture patterns and surface patterns characteristic to the burning of corpses, the only exception being the tarand grave at Põlgaste from the beginning of first millennium. Here possible secondary burial is suggested by the cremations of the bones. The colouration of cremated bone material varied within the graves. Temperatures between 600 and 800 ºC at least are suggested for funeral pyres, if the bones are of a pale colour. Nearly all graves indicated incompletely burned bone fragments of dark brown and black; the proportion was higher for the Maidla stone graves in western Estonia and for the south-eastern Estonian sand barrow cemeteries. The Põlgaste tarand grave and Uugla III stone grave both indicated relatively uniform colouration of bone material. The burial practices in west and south-eastern Estonia were different, for example in south-eastern graves no pyre sites have been found in the grave areas. In contrast, in west Estonian graves pyre sites are present and thus cremation probably took place at the burial site, with most of the cremains scattered or buried nearby. The crushing of bones before burial is not observable at south-eastern barrow cemeteries, although it is not ruled out for Late Iron Age graves. Child cremations are more frequent in south-eastern barrow cemeteries from (4th)6th–8th centuries, and in (4th)5th–6th century Maidla I stone grave in comparison with Late Iron Age graves. At the Late Iron Age Maidla II stone grave child cremation was extremely rare. The other Late Iron Age graves were too small to draw this kind of conclusion. There are many possible reasons for child under-enumeration in graves, for example the segregation of sub-adult burial is plausible, in addition to which the crushing of cremains before burial is not ruled out in Late Iron Age western Estonia. In this case the fragile cremains of infants and children become invisible in graves. Insufficient excavation technique is less plausible, as in the first Maidla grave the number of sub-adults is quite representative. The second reason why infant and child burials are rare in cremations is the taphonomic factors and/or the excavation technique. The change in mortuary practise that we observed through different bone fragmentation stages and in the proportion of determined bone units is temporal on one hand, as in the later graves with scattered cremations the average as well as median size of bone fragments is smaller. On the other hand the difference in mortuary practice, which is observable through the presence of infant and child cremations in graves, is cultural as well as temporal. More precisely, the cremains of infants and children were rarely found in Late Iron Age graves, but were often found in the graves of the Long Barrow Culture in south eastern Estonia from (4th)6th ̶ 8th centuries and well represented in the (4th)5th–6th century first stone grave

66 at Maidla. To support these hypotheses the necessity to analyse more cremated bone materials from various dates and from different cultural areas is obvious. The first stone grave at Maidla, from the Middle Iron Age in west Estonia, was used 0 by one family at most. The crude death rate was very high (e 0=10.9) and probably the community was not sustainable and inhabited the site, or used cemetery, for a short period only. In the Late Iron Age a community of 8–9 individuals – probably one family or household – used the second stone grave at Maidla. Life expectancy at birth was 25.7 years, the crude death rate was 38.9‰, and women gave birth to 3.88–5.1 children on average in the 10th–13th centuries. Life expectancy at birth shows low child mortality and favourable living conditions, or an under- representation of sub-adult (older than 5 years) burials in the grave. Concerning the community size, the results of osteological analysis of cremains from the Middle Iron Age barrow cemeteries at Rõsna (south-eastern Estonia) indicate the same pattern – one family or household commonly used one barrow cemetery. Life expectancy at birth at Rõsna during the Middle Iron Age was 17.1 years, the crude death rate was 58.5‰ on average, and women gave birth to 4.2–7.8 children, indicating relatively unfavourable living conditions in comparison with Late Iron Age Maidla. The present study demonstrates that demographic figures depend on the model or sample we choose. The overall variability in the proportion of adults and sub-adults in graves is not striking, but when skeletal samples are very small, any minor change in proportions of age cohorts has a remarkable impact on demographic figures. The acquired knowledge also reminds us to stay critical of the material we work with. It is essential to consider all the facts carefully before making palaeodemographic estimates and reconstructing life-cycles on the basis of the single graves of our ancestors. The mortuary practices and life-cycles of our ancestral communities is a puzzle that we will probably be piecing together for a long time to come. Research might be more fruitful if based on a certain region, comprised different burial places, and was fairly retrospective. These regional (bio) archaeological studies would then enable us to follow the transitions of burial customs and the possible reuse of burial grounds, and to get more reliable data for palaeodemographic predictions.

67 REFERENCES

Acsádi, G. Y., Nemeskéri, J. (1970). History of Human Life Span and Mortality. Budapest: Akadémiai Kiadó. Alesan, A., Malgosa, A., Simó, C. (1999). Looking into demography of an Iron Age population in the Western Mediterranean. I. Mortality. American Journal of Physical Anthropology, 110, 285–301. Allmäe, R. (1998). Tääksi 14.–18. sajandi populatsiooni demograafiline analüüs ja kehapikkuse rekonstrueerimine. In: V. Lang, ed., Loodus, inimene ja tehnoloogia. Interdistsiplinaarseid uurimusi arheoloogias. Muinasaja teadus, 5. Tallinn: Ajaloo Instituut, 163–187. Allmäe, R. (2003). Läänemaa 5.–13. sajandi kalmete antropoloogiline aines. In: M. Mandel, Läänemaa 5.–13. sajandi kalmed. Lisa 1. Eesti Ajaloomuuseum. Töid ajaloo alalt, 5. Tallinn: Eesti Ajaloomuuseum, 243–262. Allmäe, R. (2004). Cremations of western Estonia in 5th–13th cc. In: International Scientific Conference “200 Years of Lithuanian Anthropology: Modern Trends, History, Relation to Medical Practice and Humanities”. Vilnius, 26. Allmäe, R. (2006). Grave 2 of Maidla – the burial site of a single family. Estonian Journal of Archaeology, 10(1), 3–23. Allmäe, R. (2010). Some remarks on Kaseküla stone-cist grave, Läänemaa, Estonia. Fennoscandia Archaeologica, XXVII, 45–52. Allmäe, R. (2013). Observations on Estonian Iron Age cremations. Archaeologia Baltica, 19, 31–47. Allmäe, R. (2014a). The demography of Iron Age graves in Estonia. Lietuvos Archeologija, 40, 103–120. Allmäe, R. (2014b). Rõsna kääbaskalmistud Põhja-Setumaal: põletusmatuste uuringutulemused. In: Ü. Tamla, V. Lang, eds., Ajast ja ruumist. Uurimusi Mare Auna auks. Muinasaja teadus, 25. Tallinn; Tartu: Tallinna Ülikooli Ajaloo Instituut, Tartu Ülikooli ajaloo ja arheoloogia instituut, 39–50. Allmäe, R., Aun, M., Maldre, L. (2007a). Cremations of the Culture of Long Barrows in Northern Setumaa in the second half of the first millenium. Preliminary results. Humanbiologia Budapestinensis, 30, 113–122. Allmäe, R., Aun, M. Maldre, L. (2007b) = Аллмяэ Р., Аун М., Малдре Л. (2007). Предварительные результаты изучения остеогического материала курганных могильников Рысна-Сааре I и II в Северной Сетумаа (Юго-Восточная Эстония). In: Археология и история Пскова и Псковской земли. Семинар имени академика В. В. Седова. Материалы LII заседания, посвященного памяти профессора А. Р. Артемьева. Псков: Институт археологии РАН, 298−310. Allmäe, R, Aun. M, Maldre, L. (2008) = Аллмяэ Р., Аун М., Малдре Л. (2008). К вопросу о значении длинных курганов (по археологическим и остеологическим данным). In: Археология и история Пскова и Псковской земли. Семинар имени академика В. В. Седова. Материалы LIII заседания . Псков: Институт археологии РАН, 303−312.

68 Allmäe, R., Aun, M., Maldre, L. (2009). Rõsna-Saarõ I kääbaskalmistu. In: M. Aun, ed., Setomaa 2. Vanem ajalugu muinasajast kuni 1920. aastani. Tartu: Eesti Rahva Muuseum, 88–94. Allmäe, R., Maldre, L. (2005). Rõsna-Saare I kääbaskalmistu – esialgseid osteoloogilisi andmeid. In: Ü. Tamla, ed., Setumaa kogumik, 3. Uurimusi Setumaa loodusest, ajaloost ja folkloristikast. Tallinn: Ajaloo Instituut, 121–137. Arcini, C. (2005). Pyre sites before our eyes. In: T. Artelius, F. Svanberg, eds., Dealing with the Dead. Archaeological Perspective on Prehistoric Scandinavian Burial Ritual. Stockholm: National Heritage Board, 63–72. Aun, M. (1992) = Аун, М. (1992). Археологичеческие памятники второй половины 1-го тысячелетия н.э. в Юго-Восточной Эстонии. Таллинн: Олион. Aun, M. (2005). Pikk-kääbaste ehitusest. In: Ü. Tamla, ed. Setumaa kogumik 3. Uurimusi Setumaa loodusest, ajaloost ja folkloristikast. Tallinn: Ajaloo Instituut, 97–120. Aun, M. (2006). Pikk-kääpad Peipsi-Pihkva järve ümbruses. In: H. Valk, ed., Etnos ja kultuur: uurimusi Silvia Laulu auks. Muinasaja teadus, 18. Tartu, Tallinn, 113–128. Aun, M. (2009). Keskmine rauaaeg ja viikingiaeg (450–1050 pKr). In: M. Aun, ed., Setomaa, 2. Vanem ajalugu muinasajast kuni 1920. aastani. Tartu: Eesti Rahva Muuseum, 83–92. Aun, M., Allmäe, R., Maldre, L. (2008). Pikk-kääbaste tähendusest (Rõsna küla kääbaskalmistute materjali põhjal). In: Ü. Tamla, ed., Setumaa kogumik 4. Uurimusi Setumaa loodusest, ajaloost ja folkloristikast. Tallinn: Tallinna Ülikooli Ajaloo Instituut, 269–290. Bass, W. M. (2005). Human Osteology: A Laboratory and Field Manual. 5th edition. Missouri Archaeological Society. Becker, J. (2007). Childhood among the Etruscans: Mortuary programs at Tarquinia as indicators of the transition to adult status. Hesperia Supplements, 41: Constructions of Childhood in Ancient Greece and Italy, 281–292. Bennet, J. L. (1997). Thermal alteration of buried bone. Journal of Archaeological Science, 26, 1–8. Blumfeldt, E. (1937). Eesti majandusajalugu, I. Tartu: Akadeemiline Koperatiiv. Bocquet, J.-P., Masset, C. (1977). Estimateurs en Paléodémographie. L'Homme, XVIII, 65– 90. Bocquet-Appel, J.-P., Masset, C. (1982). Farewell to paleodemography. Journal of Human Evolution, 11, 321–333. Bronk Ramsey, C. (2009). Bayesian analysis of radiocarbon dates. Radiocarbon, 51(1), 337– 360. Brothwell, D. R. (1981). Digging up bones: The excavation, treatment and study of human skeletal remains, 3rd edition. London: Trustees of the British Museum.

69 Buckberry, J. (2000). Missing, presumed buried? Bone diagenesis and the under- representation of Anglo-Saxon children. Assemblage, 5. Available at: https://bradscholars.brad.ac.uk/handle/10454/676 [Accessed 22.09.2016]. Buikstra, J. E., Swegle, M. (1989). Bone modification due to burning: eperimental evidence. In: R. Bonnischen, M. H. Sorg, eds., Bone Modification. Orono: Peopling of the Americas Publications. Center for the Study of the First Americans, Institute for Quaternary Studies, University of Maine, 247–258. Buikstra, J. E., Ubelaker, D. H., eds. (1994). Standards for Data Collection from Human Skeletal Remains. Proceedings of a Seminar at the Field Museum of Natural History Organized by Jonathan Haas. Arkansas Archeological Survey Research Series, 44. Fayetteville: Arkansas Archeological Survey. Chamberlain, A. (2006). Demography in Archaeology. Cambridge: Cambridge University Press. De Mulder, G., Creemers, G., Van Strydonck, M. (2014). Challenging the traditional chronological framework of funerary rituals in the Meuse-Demer-Scheldt region: 14 results from the site of Lummen-Meldert (Belgium). Radiocarbon, 56(2), 461–468. De Mulder, G., Van Strydonck, M., Boudin, M. (2009). The impact of cremated bone dating on the archaeological chronology of the Low Countries. Radiocarbon, 51(2), 579–600. De Mulder, G., Van Strydonck, M., Boudin, M., Leclercq, W., Paridaens, N., Warmenbol, E. (2007). Re-evaluation of the Late Bronze Age and Early Iron Age chronology of the Western Belgian urnfields based on 14C dating of cremated bones. Radiocarbon, 49(2), 499–514. Facchini, F., Veschi, S. (2004). Age determination on long bones in a skeletal subadults sample (b-12 years). Collegium Antropologicum, 28(1), 89–98. Faerman, M., Smith, P. (2008). Has society changed its attitude to infants and children? Evidence from archaeological sites in the southern Levant. In: F. G. Jener, S. Muriel, C. R. O. Puyoles, Nasciturus, infans, puerulus vobis mater terra: la muerte en la infancia. Diputació de Castelló: Servei d'Investigacions Arqueològiques i Prehistòriques, 211–230. Available at: http://dialnet.unirioja.es/servlet/oaiart?codigo=2796721 [Accessed 14.04.2009]. Fairgive, S. I. (2008). Forensic Cremation. Recovery and Analyses. Boca Raton: CRC Press. Finlay, N. (2000). Outside of life: tradition of infant burial in Ireland from cillin to cist. World of Archaeology, 31(3), 407–422. Formisto, T. (1996). Osteological analyses. In: P. Purhonen, ed., Vainiomäki – A Merovingian Period Cemetery in Laitila, Finland. Helsinki: National Board of Antiquities, 81–87. Gejvall, N.-G. (1948). Benbestämniger. In: K. E. Salhström, N.-G. Gejvall, eds., Gravfältet på Kyrkbacken i Horns socken, Västergötland. KVHAA, 60(2), 153–180. Gejvall, N.-G. (1959). Vanligaste ben: Nagot om bestämning av brända ben och deras vetenskapliga värde. Fynd. Göteborgs och Bobus läns tidskrift, 40–47. Gejvall, N.-G. (1961). Appendix II. Anthropological and osteological analysis of the skeletal material and cremated bones from Simris 23, Simris Parish. In: B. Stjernqvist, ed., Simris II.

70 Bronze Age Problems in the Light of the Simris Excavation. Acta Archaeologica Lundensia, Series in 4–5. Lund, 157–173. Gejvall, N.-G. (1963). Cremations. In: D. Brothwell, E. Higgs, eds., Science in Archaeology. London: Thames and Hudson, 468–479. Gejvall, N.-G. (1981). Determination of burned bones from prehistoric graves. Ossa Letters, 2, Stockholm: Osteological Research Laboratory at the University of Stockholm, 7–13. Conçalves, D. (2012). Cremains. The Value of Quantitative Analysis for the Bioanthropological Research of Burned Human Skeletal Remains. PhD thesis. Coimbra: Univesity of Coimbra. Gonçalves, D., Thompson, T. J. U., Cunha, E. (2011). Implications of heat-induced changes in bone on the interpretation of funerary behaviour and practice. Journal of Archaeological Science, 38, 1308–1313. Green, M. (1998). Humans as ritual victims in the later pre-history of Western Europe. Oxford Journal of Archaeology 17(2), 169–189. Guy, H., Masset, C. Baud, C.-A. (1997). Infant taphonomy. International Journal of Osteoarchaeology, 7, 221–229. Harvig, L. (2015). Past cremation practices from a bioarchaeological perspective. How new methods and techniques revealed conceptual changes in cremation practices during the late Bronze and early Iron Age in Denmark. In: T. Thompson, ed., The Archaeology of Cremation. Burned Human Remains in Funerary Studies. Studies in Funerary Archaeology, 8. Oxford: Oxbow Books, 43–62. Harvig, L., Lynnerup, N. (2013). On the volume of cremated remains e a comparative study of archaeologically recovered cremated bone volume as measured manually and assessed by Computed Tomography and by Stereology. Journal of Archaeological Science, 40, 2713– 2722. Harvig, L., Lynnerup, N., Amsgaard Ebsen, J. (2012). Computed tomography and computed radiography of Late Bronze Age cremation urns from Denmark: An interdisciplinary attempt to develop methods applied in bioarchaeological cremation research. Archaeometry, 54(2), 369–387. Harvig, L., Runge, M. T., Lundø, M. B. (2014a). Typology and function of Late Bronze Age and Early Iron Age cremation graves – a microregional case study. Danish Journal of Archaeology. Available at: http://dx.doi.org/10.1080/21662282.2014.942980. [Accessed 17.03.2015]. Harvig, L., Frei, K. M., Price, T. D., Lynnerup, N. (2014b). Strontium isotope signals in cremated petrous portions as indicator for childhood origin. PLOS ONE, 9, 7, E101603. Available at: DOI: 10.1371/journal.pone.0101603. [Accessed 17.03.2015]. Heapost, L. (2007). The cemetery of Siksälä: osteological and paleodemographical analysis. In: S. Laul, H. Valk, Community at the Frontiers. Iron Age and Medieval. CCC Papers, 10. Tallinn; Tartu: Tartu University, 213–236. Henneberg, M. (1975). Notes on the reproduction possibilities of human prehistorical populations. Przeglad Anthropologisczny, XLI, 75–89.

71 Holck, P. (1986). Cremated Bones: A Medical-Anthropological Study of an Archaeological Material on Cremation Burials. Antropologiske skrifter, 1. Oslo: Anatomisk institutt, Universitet i Oslo. Holck, P. (1997). Why are small children so seldom found in cremations? In: E. Smits, E. Iregren, A. G. Drusini, eds., Cremation Studies in Archaeology. Proceedings of the Symposium Amsterdam 26.–27. October 1995, Amsterdam. Saonora: Logos Edizioni, 33–38. Holck, P. (1997/2008). Cremated Bones. A Medical-Anthropological Study of an Archaeological Material on Cremations Burials. Reprint of 3rd, revised edition. Oslo: University of Oslo. Hüls, C. M., Nadeau, M. J., Grootes, P. M., Erlenkeuser, H., Andersen, N. (2010). Experimental study on the origin of cremated bone apatite carbon. Radiocarbon, 52(2), 587– 599. Iregren, E. (1972). Vårby och Vårberg II: studie av kremerat människo- och jurbensmaterialm från järnåldern. Theses and Papers in North-European Archaeology. Stockholm: University of Stockholm. Iregren, E. (1983). Förhistoriska kremationer. Västmanlands Fornminnesförening och Västmanlands Läns Museums Årsskrift, 61, 23–40. Iregren, E. (1997). Why animal bones in human graves – an attempt to interpret animals present in Iron Age cremations in Sweden. In: E Smits, E Iregren, A. G. Drusini, eds., Cremation Studies in Archaeology. Proceedings of the Symposium Amsterdam 26.–27. October 1995, Amsterdam. Saonora: Logos Edizioni, 9–32. Jaanits, L., Laul, S., Lõugas, V., Tõnisson, E. (1982). Eesti esiajalugu. Tallinn: Valgus. Jaeger, J. H., Johansen, V. L. (2013). The cremation of infants/small children: An archaeological experiment concerning the effects of fire on bone weight. Cadernos do GEEvH, 2(2), 13–26. Jankauskas, R. (2002). Anthropology of the Iron Age inhabitants of Lithuania. In: P. Bennike, É. B. Bodszár, C. Susanne, eds., Ecological Aspects of Past Human Settlements in Europe. Biennal Books of European Association of Anthropologists, 2. Budapest: Eötvös University Press, 129–142. Jankauskas, R. (2009). Anthropologische Bestimmungen. Skeletal inventory, age, sex, withers height and pathologies of Marvele sample. In: M. Bertašius, Marvelė, Ein Bestattungsplatz mit Mittellitauischen Pferdegräbern, II Band. Kaunas: Kauno technologijos universitetas, 93–98. Jankauskas, R., Urbanavicius, A. (1998). Palaeodemography of Marvele. Acta Biologica Szegediensis, 42, 67–73. Jonuks, T. (2009). Eesti muinasusund. Dissertationes archaeologiae Universitas Tartuensis, 2. Tartu: Tartu Ülikooli Kirjastus. Kaliff, A. (1992). Brandgravskick och föreställningsvärld. En religionsarkeologisk diskussion. Occasional Papers in Archaeology, 4. Uppsala: Societas Archaeologica Upsaliensis. Kaliff, A. (1997). Grav oh kultplats. Eskatologiska föreställingar under yngrebronsalder och alder järnalder i Östergotland. Aun, 24. Uppsala: Uppsala University.

72 Kaliff, A., Oestigaard, T. (2004). Cultivating corpses. A comparative approach to disembodied mortuary remains. Current Swedish Archaeology, 12, 83–104. Kalling, K. (1993). Viimsi kalmete luuainese antropoloogiline analüüs. In: V. Lang, ed., Kaks tarandkalmet Viimsis, Jõelähtme kihelkonnas. Lisa 1. Tallinn: Eesti Teaduste Akadeemia Ajaloo Instituut, 67–69. Kalling, K. (1995). Paleoantropoloogilisi andmeid Tartu Jaani kiriku kalmistu 13.–14. sajandi matuste kohta. In: H. Valk, ed., Tartu arheoloogiast ja vanemast ehitusloost. Tartu Ülikooli Arheoloogiakabneti Toimetised, 8. Tartu: Tartu Ülikooli Arheoloogia kabinet, 47– 57. Kalling, K. (1997). Uusi paleoantropoloogilisi andmeid Tartu Jaani kiriku kalmistu kohta. In: H. Valk, ed., Arheoloogilisi uurimusi, I. Tartu Ülikooli Arheoloogiakabneti Toimetised, 9. Tartu: Tartu Ülikooli Arheoloogia kabinet, 54–70. Kalman, J. (2000a). Stone grave II of Tõugu – skeletal report. In: V. Lang, Keskusest ääremaaks. Viljelusmajandusliku asustuse kujunemine ja areng Vihasoo–Palmse piirkonnas Virumaal. Muinasaja teadus, 7. Tallinn: Ajaloo Instituut, 387–402. Kalman, J. (2000b). Tandemägi stone grave – osteological report. In: V. Lang, Keskusest ääremaaks. Viljelusmajandusliku asustuse kujunemine ja areng Vihasoo–Palmse piirkonnas Virumaal. Muinasaja teadus, 7. Tallinn: Ajaloo Instituut, 423–434. Kalman, J. (2000c). Uusküla II skeletal analysis. In: V. Lang, Keskusest ääremaaks. Viljelusmajandusliku asustuse kujunemine ja areng Vihasoo–Palmse piirkonnas Virumaal. Muinasaja teadus, 7. Tallinn: Ajaloo Instituut, 437–440. Kalman, J. (2000d). Skeletal analysis of the graves of Kaseküla, Poanse I and Poanse II. In: T. Tamla, ed., Eesti Ajaloomuusem. Töid ajaloo alalt, II. Tallinn: Eesti Ajaloomuuseum, 17– 40. Khvoshchinskaya, N. V. (2004) = Хвощинская, Н. В. (2004). Финны на западе Новгородской земли (По материалaм могильника Залахтовье). C.-Петербург: Институт Истории Материальной Культуры. Kiristaja, A. (2009). Kiviaeg, pronksiaeg ja vanem rauaaeg. In: M. Aun, ed., Setomaa, 2. Vanem ajalugu muinasajast kuni 1920. aastani. Tartu: Eesti Rahva Muuseum, 23–67. Kivirüüt, A. (2011). Põletatud luude uurimine: metoodika ning praktika Viimsi I tarandkalme leiukompleksi näitel. Bakalaureusetöö. Tartu: Tartu Ülikool. Available at: http://www.arheo.ut.ee/docs/BA11_Kivirüüt.pdf [Accessed 24.05.2013] Kivirüüt, A. (2014). A Comparative Osteological and Intra-site Spatial Analysis of Tarand- graves. MA dissertation. Tartu: University of Tartu. Available at: http://dspace.ut.ee/handle/10062/41912 [Accessed 24.04.2016] Konsa, M. (2013). Intrasite spatial analysis of the cemeteries with dispersed cremation burials. In: E. Graeme et al, eds., Archaeology in the Digital Era, II: 40th Conference on Computer Applications and Quantitative Methods in Archaeology, Southampton, 26–30 March 2012. Amsterdam: Amsterdam University Press, 570−574. Konsa, M., Engbring, L. (2011). Analysis of dispersed cremations. In: R. Allmäe, L. Heapost, J. Limbo-Simovart, eds., Programme and Abstracts: Baltic Bioarchaeology

73 Meeting. Past People around the Baltic Sea. August 21–24, 2011, Tallinn, Estonia. Tallinn: Tallinna Ülikooli Ajaloo Instituut, 19. Kurila, L. (2009). The Social Organisation in East Lithuania in the 3rd–12th Centuries (on the Basis of the Mortuary Record) (Summary of doctoral dissertation). Vilnius: Vilnius University. Available at: http://vddb.library.lt/fedora/get/LT-eLABa- 0001:E.02~2009~D_20091109_151308-59594/DS.005.1.01.ETD [Accessed 23.11.2013]. Kurila, L. (2015a). The accuracy of the osteological sexing of cremated human remains: A test based on grave goods from East Lithuanian barrows. Collegium Antropologicum, 39(4), 821–828. Kurila, L. (2015b). Social classes in the Iron Age east Lithuania: An attempt of identification in the mortuary record. In: A. Haak, V. Lang, M: Lavento, eds., Today I Am Not the One I Was Yesterday: Archaeology, Identity, and Change. Interarchaeologia, 4. Tartu: University of Tartu, 45–63. Kurila, L. (2015c). Cremation as limitation? A paleodemographic inquiry into the accuracy of macroscopic analysis of cremated human remains based on an East Lithuanian sample. Papers on Anthropology, XXIV(2), 67–85. Lahtiperä, P. (1970). Luuaineston analyysi. In: U. Salo, ed., Metallikautinen asutus Kokemäenjoen suussa, II. Pori: Satakunnan Museon kannatusyhdistys, 198–219. Laneman, M. (2012). Stone-cist grave at Kaseküla, Western Estonia, in the light of AMS dates of human bones. Estonian Journal of Archaeology, 16(2), 91–117. Laneman, M., Lang, V. (2013). New radiocarbon dates for two stone-cist graves at Muuksi, northern Estonia. Estonian Journal of Archaeology, 17(2), 89–122. Laneman, M., Lang, V., Malve, M., Rannamäe, E. (2015). New data on Jaani stone graves at Väo, northern Estonia. Estonian Journal of Archaeology, 19(2), 110–137. Lang, V. (1993). Kaks tarandkalmet Viimsis Jõelähtme kihelkonnas. Tallinn: Eesti Teaduste Akadeemia Ajaloo Instituut. Lang, V. (1996). Muistne Rävala. Muistised, kronoloogia ja maaviljelusliku asustuse kujunemine Loode-Eestis, eriti Pirita jõe alamjooksu piirkonnas, I–II. Muinasaja teadus, 4. Tallinn: Eesti Teaduste Akadeemia Ajaloo Instituut. Lang, V. (2000). Keskusest ääremaaks. Viljelusmajandusliku asustuse kujunemine ja areng Vihasoo-Palmse piirkonnas Virumaal. Muinasaja teadus, 7. Tallinn: Ajaloo Instituut. Lang, V. (2006). Varased tarandkalmed Eestis. In: H. Valk, ed., Etnos ja kultuur. Uurimusi Silvia Laulu auks. Muinasaja teadus, 18. Tartu; Tallinn: Tartu Ülikooli arheoloogia õppetool, Tallinna Ülikooli Ajaloo Instituut, 53–78. Lang, V. (2007a). The Bronze and Early Iron Ages in Estonia. Estonian Archaeology, 3. Tartu: Tartu University Press. Lang, V. (2007b). Baltimaade pronksi- ja rauaaeg. Tartu: Tartu Ülikooli Kirjastus. Lang, V., Laneman, M., Ilves, K., Kalman, J. (2001). Fossil fields and stone-cist graves of Rebala revisited. Archaeological Fieldwork in Estonia / Arheoloogilised välitööd Eestis 2000, 34–47.

74 Lang, V., Ligi, P. (1991). Muistsed kalmed ajaloolise demograafia allikana. In: L. Jaanits, V. Lang, eds., Arheoloogiline kogumik. Muinasaja teadus, 1. Tallinn: Eesti Arheoloogiaselts; Eesti Teaduste Akadeemia Ajaloo Instituut, 216–238. Lange, M., Schutkowski, H., Hummel, S., Herrmann, B. (1987). A Bibliography on Cremation. PACT, 19. Rixensart: Hackens. Lanting, J. N., Aerts-Bijma, A. T., van der Plicht, J. (2001). Dating of cremated bones. Radiocarbon, 43, 249–254. Larsson, A. M., Nilsson Stutz, L. (2014). Reconcilable differences: cremation, fragmentation and inhumation in Mesolithic and Neolithic Sweden. In: I. Kuijut, P. Q. Quinn, G. Cooney, eds., Transformation by Fire. The Archaeology of Cremation in Cultural Context. Arizona: University of Arizona Press, 189–206. Laul, S. (2001). Rauaaja kultuuri kujunemine Eesti kaguosas (500 e.Kr.–500 p.Kr.). Muinasaja teadus, 9 / Õpetatud Eesti Seltsi Kirjad, 7. Tallinn: Ajaloo Instituut, Õpetatud Eesti Selts. Ligi, H. (1961). Eesti talurahva olukord ja klassivõitlus Liivi sõja algul (1558–1561). Tallinn: ENSV Teaduste Akadeemia. Ligi, P. (1989). = Лиги, П. (1989). Культура длинных курганов в свете данных палеодемографии. Eesti Teaduste Akadeemia Toimetised, Ühiskonnateadused, 38(4), 316– 322. Lillak, A. (2009). Maa-alused põletusmatused Setomaal. In: M. Aun, ed., Setomaa, 2. Vanem ajalugu muinasajast kuni 1920. aastani. Tartu: Eesti Rahva Muuseum, 82. Lindquist, M. 1981. Mylingar offer – utsatta barn eller förhistorika barnegravingsar? Gotländskt Arkiv, 53, 7–12. Lovejoy, C. O., Russell, K. F., Harrison, M. L. (1990). Long bone growth velocity in Libben Population. American Journal of Human Biology, 2, 533–541. Mägi, M. (2002). At the Crossroads of Space and Time. Graves, Changing Society and Ideology on Saaremaa (Ösel), 9th–13th Centuries AD. CCC papers, 6. Tallinn: Institute of History, Tallinn, Gotland University College, Centre of Baltic Studies. Mägi, M. (2004). Changing trends in the study of Estonian Iron Age cemeteries: archaeological excavations in archives. In: G. Carver, ed., Digging in the Dirt: Excavation in a New Millennium. British Archaeological Reports, International Series, S1256. Oxford: Hadrian Books, 11−18. Mägi, M. (2005a). Mortuary houses in Iron Age Estonia. Estonian Journal of Archaeology, 9(2), 93−131. Mägi, M. (2005b). Kivikalmed ja surnumajad. Matmiskombestiku areng rauaaegsel Saaremaal. Saaremaa Muuseum. Kaheaastaraamat, 2003–2004. Kuressaare: Saaremaa Muuseum, 3−34. Mägi, M. (2007). Collectivity versus individuality: the warrior ideology of Iron Age burial rites on Saaremaa. Archaeologia Baltica, 8, 263−272. Mägi, M. (2013). Late prehistoric societies and burials in the Eastern Baltic. Archaeologia Baltica, 19, 177−194.

75 Mägi, M., Allmäe, R., Maldre, L. (1998). Viking Age graveyard at Piila, Saaremaa. Archaeological Fieldwork in Estonia / Arheoloogilised välitööd Eestis 1997, 99–116. Maldre, L. (2003). Läänemaa kivikalmte arheozooloogiline materjal. In: M. Mandel, Läänemaa 5.–13. sajandi kalmed. Lisa 1. Eesti Ajaloomuuseum. Töid ajaloo alalt, 5. Tallinn: Eesti Ajaloomuuseum, 243–263. Mandel, M. (1973). Aruanne Kaseküla kivikirstkalme kaevamistest 1973. aastal. (Manuscript in the archives of the Estonian History Museum.) Mandel, M. (1975). Ausgrabungen der Steingräber von Kaseküla. Eesti NSV Teaduste Akadeemia Toimetised, Ühiskonnateadused, 24(1), 74–76. Mandel, M. (2000). Poanse tarandkalmed. In: T. Tamla, ed., Eesti Ajaloomuusem. Töid ajaloo alalt, II. Tallinn, Eesti Ajaloomuuseum, 89–112. Mandel, M. (2003). Läänemaa 5.–13. sajandi kalmed. Eesti Ajaloomuuseum. Töid ajaloo alalt, 5. Tallinn: Eesti Ajaloomuuseum. Mandel, M. (2011). Das unterirdische Grabfeld mit Feuerbestattungen zu Uugla. Archaeological Fieldworks in Estonia / Arheoloogilised välitööd Eestis 2010, 175–178. Mandel, M., Allmäe, R. (2009). Ergebnisse Der Archäologischen Ausgrabungen in Uugla. Archaeological Fieldworks in Estonia / Arheoloogilised välitööd Eestis 2008, 114–123. Mandel, M., Allmäe, R. (2013). Forschungs- und Herrichtungsarbeiten im Gräberfeld zu Maidla. Archaeological Fieldworks in Estonia / Arheoloogilised välitööd Eestis 2012, 281– 288. Mandel, M., Allmäe, R., Maldre, L. (2015). Eine weitere Überraschung vom Gräberfeld zu Maidla. Archaeological Fieldworks in Estonia / Arheoloogilised välitööd Eestis 2014, 107– 112. Mäntyla-Asplund, S., Storå, J. (2010). On the archaeology and osteology of the Rikala cremation cemetery in Salo, SW Finland. Fennoscandia Archaeologica, XXVII, 53–68. Mayne Correia, P. M. (1997). Fire modification of bone: A review of literature. In: W. D. Haglund, M. H. Sorg, eds., Forensic Taphonomy. The Postmortem Fate of Human Remains. Boca Raton: CRC Press, 275–290. Mays, S. (2006). The Archaeology of Human Bones. London; New York: Routledge. McCaa, R. (1998). Calibrating paleodemography: The uniformitarian challenge turned. American Association of Physical Anthropology Annual Meeting, April 2, 1998, Salt Lake City. Available at: http://www.hist.umn.edu/~rmccaa/paleo98/index0.htm [Accessed 12.11.2013]. McCaa, R. (2000). Paleodemography of the Americas. In: R. H Steckel, J.C. Rose, eds., The Backbone of History: Health and Nutrition in the Western Hemisphere. New York: Cambridge University Press, 94–126. McCutcheon, P. (1992). Burned archaeological bone. In: J. K. Stein, ed., Deciphering a Shell Midden. San Diego: Academic Press, 347–368. McKinley, J. (1989). Cremations: expectations, methodologies and realities. In: C. A. Roberts, F. Lee, J. Bintliff, eds., Burial Archaeology: Current Research, Methods and

76 Developments. British Archaeological Reports, British Series, 211. Oxford: British Archaeological Reports, 65–76. McKinley, J. (1994a). Bone fragment size in British cremation burials and its implication for pyre technology and ritual. Journal of Archaeological Science, 21(3), 339–342. McKinley, J. (1994b). A pyre and grave goods in British cremation burials; Have we missed something? Antiquity, 68, 132–134. McKinley, J. (1994c). The Anglo-Saxon Cemetery at Spong Hill, North Elmham, Part VIII: The Cremations. East Anglian Archaeology, 69. Available at: http://eaareports.org.uk/publication/report69/ [Accessed 30.04.2016]. McKinley, J. I. (2008). In the heat of the pyre: efficiency of oxidation in Romano-British cremations – did it really matter? In: C. W. Schmidt, S. A. Symes, eds., The Analysis of Burned Human Remains. London: Academic Press, 162–184. McKinley, J. I., Bond, J. M. (2001). Cremated bone. In: D. R. Brothwell, A. M. Pollard, eds., Handbook of Archaeological Sciences. Chicester: Wiley, 281–292. Miles, A. E. W. (1963). Dentition in the assessment of individual age in skeletal material. In: D. R. Brothwell, ed., Dental Anthropology. New York: Pergamon Press, 191–209. Molleson, T. (1991). Demographic implications of the age structure of early English cemetery samples. Actes des Jornees Anthropologiques, 5, 113–121. Musgrave, J., Prag, A. J. N. W., Neave, R., Fox, R. L., White, H. (2010). The occupants of tomb II at Vergina. Why Arrhidaios and Eurydice must be excluded. International Journal of Medical Science, 7(6), 1–15. Naysmith, P., Scott, E. M., Cook, G. T., Heinemeier, J., van der Plicht, J., Van Strydonck, M., Bronk Ramsey, C., Grootes, P. M., Freeman, S. P. H. T. (2007). Cremated bone intercomparison study. Radiocarbon, 49(2), 403–408. Olsen, J., Heinemeier, J., Hornstrup, K. M., Bennike, P., Thrane, H., (2013). “Old wood” effect in radiocarbon dating of prehistoric cremated bones? Journal of Archaeological Science, 40, 30–34. Palli, H. (1996). Eesti rahvastiku ajalugu aastani 1712. Academia, 6. Tallinn: Teaduste Akadeemia Kirjastus. Pankowská, A., Průchová, E., Moník, M., Martina Nováková, M. (2014). Taphonomy of cremation burials: excavation and deposition bias in bone preservation. In: O. Chvojka, M. Chytráček, H. Gruber, L. Husty, J. Michálek, R. Sandner, K. Schmotz, S. Traxler, eds., Fines Transire. Archäologische Arbeitsgemeinschaft Ostbayern/West- und Südböhmen/ Oberösterreich. 23. Treffen vom 19. bis 22. Juni 2013 in Kostenz. Rahden/Westf.: Leidorf, 223–232. Parker Pearson, M. (1999). The Archaeology of Death and Burial. Sutton: Phoenix Mill. Piga, G., Malgosa, A., Mazzarello, V., Bandiera, P., Melis, P., Enzo, S. (2008). Anthropological and physicochemical investigation of the burnt remains of Tomb IX in the ‘Sa Figu’ Hypogeal Necropolis (Sassari, Italy) – Early Bronze Age. International Journal of Osteoarchaeology, 18(2), 167–177.

77 Purhonen, P., ed. (1996). Vainionmäki – A Merovingian Period Cemetery in Laitila, Finland. Helsinki: National Board of Antiquities. Purhonen, P. (1998). Kristinukon saapumisesta Suomeen. Uskontoarkeologinen tutkimus. Suomen Muinasmuistoyhdistyksen Aikakauskirja, 106. Helsinki: Suomen Muinasmuistoyhdistys. Purhonen, P. (1999). “Suku käski suolle viedä ...”: Mitä tapahtui ei-toivuille lapsille? Tieteessä tapahtuu, 17(3, 4). Available at: http://www.tieteessatapahtuu.fi/993/purhonen.htm [Accessed 19.09.2009]. Pusch, C. M., Broghammer, M., Scholz, M. (2000). Cremation practices and the survival of ancient DNA: burnt bone analyses via RAPD-mediated PCR. Anthropologischer Anzeiger, 58(3), 237–251. Reimer, P.J., Baillie, M. G. L., Bard, E., Bayliss, A., Beck, J. W., Blackwell, P. G., Bronk Ramsey, C., Buck, C. E., Burr, G. S., Edwards, R. L., Friedrich, M., Grootes, P. M., Guilderson, T. P., Hajdas, I., Heaton, T. J., Hogg, A. G., Hughen, K. A., Kaiser, K. F., Kromer, B., McCormac, F. G., Manning, S. W., Reimer, R. W., Richards, D. A., Southon, J. R., Talamo, S., Turney, C. S. M., van der Plicht, J., Weyhenmeyer, C. E. (2009). IntCal09 and Marine09 Radiocarbon Age Calibration Curves, 0–50,000 Years cal BP. Radiocarbon, 51, 1111–1150. Roberts, C., Manchester, K. (1995). The Archaeology of Disease. 2nd edition. Stroud: Sutton Publishing. Rösing, F. W., Jankauskas, R. (1997). Infant deficit in pre-modern burial sites. In: M. Thetloff, ed., The 8th Tartu international anthropological conference. 12–16 October, Tartu, Estonia. Dedicated to the 100th Birth Anniversary of Prof. Juhan Aul. Tartu: University of Tartu, Centre for Physical Anthropology, 50–52. Saunders, S. R. (1992). Subadult skeletons and growth related studies. In: S. R. Saunders, A. Katzenberg, eds., Skeletal Biology of Past Peoples: Research Methods. Chichester: Wiley, 1–20. Sigvallius, B. (1994). Funeral Pyres. Iron Age Cremations in North Spånga. Theses and Papers in Osteology, I. Stockholm: Stockholm University, Department of Archaeology, Osteological Research Laboratory. Shipman, P., Forster, D., Schoeniger, M. (1984). Burnt bones and teeth: an experimental study of colour, morphology, crystal structure and shrinkage. Journal of Archaeological Sciences, 11, 307–325. Smith, P., Avishai, G., Greene, J. A., Stager, L. E. (2011). Aging cremated infants: the problem of sacrifice at the Tophet of Carthage. Antiquity, 85, 859–874. Snoeck, C., Lee-Thorp, J., Shculting, R., De Yong, J., Debouge, W., Mattielli, N. (2015). Calcined bone provides a reliable substrate for strontium isotope ratios as shown by an enrichment experiment. Rapid Communication in Mass Spectrometry, 29(1), 107–114. Soames, J. V., Southam, J. C. (1993). Oral Pathology. 2nd edition. Oxford: Oxford University Press. Sokolovski, V. (1990) = Соколовский, В. (1990). Средневековое деревенское кладбище в Тяекси. Eesti Teaduste Akadeemia Toimetised. Ühiskonnateadused, 39(4), 414–422.

78 Stiner, M. C., Kuhn, S. L., Weiner, S., Bar-Yosef, O. (1995). Differential burning, recrystallization, and fragmentation of archaeological bone. Journal of Archaeological Science, 22, 223–237. Symes, S. A., Rainwater, C. W., Chapman, E. N., Gipson, D. R., Piper, A. L. (2008). Patterned thermal destruction of human remains in a forensic setting. In: C. W. Schmidt, S. A. Symes, eds., The Analysis of Burned Human Remains. London: Academic Press, 15–54. Tarvel, E. (1972). Adramaa. Eesti talurahva maakasutuse ja maksustuse alused 13.–19. sajandil. Tallinn: Eesti NSV Teaduste Akadeemia Ajaloo Instituut. Tvauri, A. (2012). The Migration Period, Pre-Viking Age, and Viking Age in Estonia. Estonian Archaeology, 4. Tartu: Tartu University Press. Ubelaker, D. H. (1989). Human Skeletal Remains: Excavation, Analysis, Interpretation. Manuals on Archeology, 2. 2nd edition. Washington: Taraxacum. Valk, H., Allmäe, R. (2009). Põletusmatused Siksälä Kerigumäel. In: M. Aun, ed., Setomaa, 2. Vanem ajalugu muinasajast kuni 1920. aastani. Tartu: Eesti Rahva Muuseum, 387. Valk, H., Allmäe, R. (2010). Kirikumägi at Siksälä: evidence of a new grave form of South- Eastern Estonia. Estonian Journal of Archaeology, 14(1), 40–55. Vedru, G. 2009. Layers of landscape, layers of site. Estonian Journal of Archaeology 13(1), 21–35. Veselka, B., Lemmers, S. (2014). Deliberate selective deposition of Iron Age cremations from Oosterhout (prov. Noord-Brabant, the Netherlands): a ‘pars pro toto’ burial ritual. LUNULA. Archaeologia protohistorica, XXII, 151–158. Wahl, J. (2008). Investigations on Pre-Roman and Roman cremation remains from southwestern Germany: results, potentialities and limits. In: C. W. Schmidt, S. A. Symes, eds., The Analysis of Burned Human Remains. London: Academic Press, 145–162. Walker, P. L., Johnson, J. R., Lambert, P. M. (1988). Age and sex biases in the preservation of human skeletal remains. American Journal of Physical Anthropology, 76(2), 183–188. Walker, P. L., Miller, K. P. (2005). Time, temperature, and oxygen availability: An experimental study of the effects of environmental conditions on the colour and organic content of cremated bone. American Journal of Physical Anthropology, 126(S40), 222. Walker, P. L., Miller, K. W. P., Richman, R. (2008). Time, temperature and oxygen availability: An experimental study of the effects of environmental conditions on the color and organic content of cremated bone. In: Schmidt C. W., Symes S. A. (eds.). The Analysis of Burned Human Remains. London: Academic Press, 129–137. Warren, M. W., Maples, W. R. (1997). The anthropometry of contemporary commercial cremation. Journal of Forensic Sciences, 42(3), 417–23. Waterhouse, K. (2013). The effect of victim age on burnt bone fragmentation: Implications for remains recovery. Forensic Science International, 231, 409.e1–409.e7. Welinder, S. (1998). The cultural construction of childhood in Scandinavia, 3500 BC–1350 AD. Current Swedish Archaeology, 6, 185–204.

79 Wessmann, A. (2010). Death, Destruction and Commemoration. Tracing ritual activities in Finnish Iron Age cemeteris (AD 550–1150). Iskos 18. Helsinki: The Finnish Antiquarian Society. White, T. D., Folkens, P. A. (2000). Human Osteology. San Diego: Academic Press. Whyte, T. R. (2001). Distinguishing remains of human cremations from burned animal bones. Journal of Field Archaeology, 28, 437–448. Wickholm, A. (2007). Commemorating the past. A case of collective remembering from Alsätra cemetery in Karjaa, Finland. In: A. Merkevicius, ed., Colours of Archaeology. Material Culture and Society. Interarchaeologia, 2. Vilnius: University of Vilnius, 107– 116. Wickholm, A. (2008). Reuse in Finnish cremation cemeteries under level ground. In: F. Falström, T. Oestigaard, eds., Examples of Collective Memory. Materiality of Death: Bodies, Burials, Beliefs. British Archaeological Reports, International Series, 1768. Oxford: British Archaeological Reports, 89–97. Wickholm, A., Raninen, S. (2003). Rautakautinen riesa – polttokenttäkalmistojen problematiikkaa. Muinaistutkija, 2003(2), 2–14. Wickholm, A., Raninen, S. (2006). The broken people: deconstruction of personhood in Iron Age Finland. Estonian Journal of Archaeology, 10(2), 150–166. Williams, H. (2008). Towards an archaeology of cremation. In: C. W. Schmidt, S. A. Symes, eds., The Analysis of Burned Human Remains. London: Academic Press, 239–269. Workshop (1980). = Workshop of European Anthropologists. Recommendations for age and sex diagnoses of skeletons. Journal of Human Evolution, 9(7), 517–549. Wood, J. W. (1990). Fertility in anthropological populations. Annual Review of Anthropology, 19, 211–242. Zariņa, G. (2006). The main trends in the palaeodemography of the 7th–18th century population of Latvia. Anthropologischer Anzeiger, 64, 189–202. Zariņa, G. (2009). Latvijas iedzīvotāju paleodemogrāfija: 7. g.t. pr. Kr.–1800. g. Rīga: Latvijas Vēstures institūta apgāds.

80 APPENDIX 1. TABLE 6. COMMUNITY SIZE CALCULATIONS. MODELS Life Number Grave usage in years and respective community size expectancy of Grave at birth in burials 350 250 150 100 50 yrs** Poanse I 22.3 2.9 4.1 6.8 10.2 20.5 46 tarand grave* 10.8 1.4 2.0 3.3 5.0 9.9

Poanse II 25.1 2.5 3.5 5.9 8.8 17.6 35 tarand grave* 14.4 1.4 2.0 3.4 5.0 10.1

Maidla I 30.3 1.7 2.4 4.0 6.1 12.1 20 stone grave 10.9 0.6 0.9 1.5 2.2 4.4

Maidla II 28.2 6.0 8.3 13.9 20.8 41.7 74 stone grave 27.6 5.8 8.2 13.6 20.4 40.8

Rõsna-Saare I 27.2 4.7 6.6 11.1 16.6 33,2 sand-barrow 61 cemetery 18.1 3.2 4.4 7.4 11.0 22.1

Rõsna-Saare II 26.5 3.5 4.9 8.1 12.2 24.4 sand-barrow 46 cemetery 16.9 2.2 3.1 5.2 7.8 15.5

Suure Rõsna 32.1 3.7 5.1 8.6 12.8 25.7 sand-barrow 40 cemetery 16.7 19 2.7 4.5 6.7 13.4 * Poanse demographic data from Kalman (2000d) **Life expectancy at birth from 1) non-corrected life tables and 2) after Boquet and Masset formula (1982).

81 APPENDIX 2. MEASUREMENTS OF CREMATED BONE FRAGMENTS. TABLES 7.1.-7.2.

Table 7.1. Measurements of the longest bone fragments in cremated bone units (cm). Rõsna- Rõsna- Suure- Kiriku- Sample Kirbla Ehmja Maidla I Maidla II Uugla III Põlgaste Saare I Saare II Rõsna mägi 5-7 Date 11-12 5-6 10-13 11-13 3-5 6-8 7-8 6-8 10-12 11-12 N 51 151 205 540 316 54 106 61 140 29 Ave 3.2 2.2 4.3 3.2 2.7 5.7 4.7 4.9 4.3 3.0 Mode 3.0 2.0 3.5 2.2 2.7 4.5 4.4 3.8 5.1 1.3 Median 3.0 2.0 3.8 3.0 2.7 5.4 4.4 4.6 4.0 2.4 Max 5.0 8.6 10.1 17.2 5.2 12.6 9.5 9.8 8.8 6.6 Min 1.7 0.9 1.7 1.1 0.8 1.8 1.8 1.7 1.2 1.1 STD 0.8 1.0 1.5 1.3 0.8 2.6 1.8 1.8 1.6 1.6 Var 0.693584 0.933262 2.312259 1.686081 0.657928 6.693697 3.153653 3.142858 2.658993 2.654360

Table 7.2. Measurements of the biggest cranial fragments in cremated bone units (cm*cm). Rõsna- Rõsna- Suure- Kiriku- Sample Kirbla Ehmja Maidla I Maidla II Uugla III Põlgaste Saare I Saare II Rõsna mägi 5-7 Date 11-12 5-6 10-13 11-13 3-5 6-8 7-8 6-8 10-12 11-12 N 27 75 139 248 192 47 65 40 84 15 Ave 3.4 2.7 7.4 5.0 3.2 12.9 8.6 8.9 7.3 4.6 Mode 8.6 1.0 7.1 2.2 2.4 5.5 8.4 n/a 1.9 4.6 Median 2.9 2.0 5.9 4.1 2.9 12.0 7.8 6.9 6.1 4.6 Max 8.6 15.8 28.1 32.7 11.3 31.3 23.4 42.3 23.6 11.0 Min 0.9 0.5 1.1 0.9 0.7 3.6 1.7 2.1 1.5 1.3 STD 2.3 2.7 5.0 3.8 1.5 6.4 5.2 7.6 4.8 2.7 Var 5.147379 7.108258 24.719200 14.489692 2.103583 40.603612 26.548999 58.072666 22.831203 7.284335

82 APPENDIX 3. LIFETABLES NON-CORRECTED. TABLES 8.1.-8.6. Table 8.1. Rõsna-Saare I barrow cemetery. Non-corrected lifetable. x Dx dx lx qx Lx Tx ex 0 5.00 8.20 100.00 0.08 95.90 2727.37 27.27 1-4 6.00 9.84 91.80 0.11 347.54 2631.47 28.66 5-9 6.25 10.25 81.97 0.13 384.22 2283.93 27.86 10-14 3.75 6.15 71.72 0.09 343.24 1899.70 26.49 15-19 5.96 9.78 65.57 0.15 303.43 1556.47 23.74 20-39 15.86 26.00 55.80 0.47 855.92 1253.04 22.46 40-59 14.14 23.18 29.79 0.78 364.06 397.12 13.33 60+ 4.03 6.61 6.61 1.00 33.06 33.06 5.00 61.00 100.00 2727.37

Table 8.2. Rõsna-Saare II barrow cemetery. Non-corrected lifetable. x Dx dx lx qx Lx Tx ex 0 5.00 10.87 100.00 0.11 94.57 2653.89 26.54 1-4 5.33 11.59 89.13 0.13 333.33 2559.33 28.71 5-9 3.96 8.61 77.54 0.11 366.17 2226.00 28.71 10-14 3.71 8.06 68.93 0.12 324.50 1859.83 26.98 15-19 2.94 6.40 60.87 0.11 288.35 1535.33 25.22 20-39 11.78 25.60 54.47 0.47 833.33 1246.98 22.89 40-59 9.44 20.53 28.86 0.71 371.98 413.65 14.33 60+ 3.83 8.33 8.33 1.00 41.67 41.67 5.00 46.00 100.00 2653.89

Table 8.3. Suure Rõsna barrow cemetery. Non corrected lifetables. x Dx dx lx qx Lx Tx ex 0 3.00 7.50 100.00 0.08 96.25 3208.29 32.08 1-4 4.42 11.04 92.50 0.12 347.92 3112.04 33.64 5-9 4.79 11.98 81.46 0.15 377.34 2764.12 33.93 10-14 2.79 6.98 69.48 0.10 329.95 2386.78 34.35 15-19 0.50 1.25 62.50 0.02 309.38 2056.83 32.91 20-39 6.17 15.43 61.25 0.25 1070.74 1747.45 28.53 40-59 12.50 31.26 45.82 0.68 603.89 676.71 14.77 60+ 5.83 14.56 14.56 1.00 72.82 72.82 5.00 40.00 100.00 3208.29

83 Table 8.4. Sand barrow cemeteries at Rõsna (summarised data). Non-corrected lifetables. x Dx dx lx qx Lx Tx ex 0 13.00 8.84 100.00 0.09 95.58 2835.24 28.35 1-4 15.75 10.71 91.16 0.12 343.20 2739.66 30.05 5-9 15.00 10.20 80.44 0.13 376.70 2396.46 29.79 10-14 10.25 6.97 70.24 0.10 333.76 2019.76 28.76 15-19 9.41 6.40 63.27 0.10 300.33 1686.00 26.65 20-39 33.81 23.00 56.87 0.40 907.31 1385.68 24.37 40-59 36.09 24.55 33.86 0.72 431.80 478.37 14.13 60+ 13.69 9.31 9.31 1.00 46.57 46.57 5.00 147.00 100.00 2835.24 Table 8.5. Maidla I stone grave. Non-corrected life table. x Dx dx lx qx Lx Tx ex 0 0.00 0.00 100.00 0.00 100.00 3064.68 30.65 1-4 3.00 15.00 100.00 0.15 370.00 2964.68 29.65 5-9 3.00 15.00 85.00 0.18 387.50 2594.68 30.53 10-14 2.00 10.00 70.00 0.14 325.00 2207.18 31.53 15-19 0.50 2.50 60.00 0.04 293.75 1882.18 31.37 20-39 2.90 14.48 57.50 0.25 1005.19 1588.43 27.62 40-59 6.56 32.81 43.02 0.76 532.22 583.24 13.56 60+ 2.04 10.20 10.20 1.00 51.02 51.02 5.00 20.00 100.00 3064.68 Table 8.6. Maidla II stone grave. Non-corrected life table. x Dx dx lx qx Lx Tx ex 0 7.00 9.46 100.00 0.09 95.27 2815.95 28.16 1-4 5.83 7.88 90.54 0.09 346.40 2720.68 30.05 5-9 8.17 11.04 82.66 0.13 385.70 2374.28 28.72 10-14 1.00 1.35 71.62 0.02 354.74 1988.58 27.76 15-19 1.81 2.45 70.27 0.03 345.23 1633.84 23.25 20-24 6.70 9.06 67.82 0.13 316.45 1288.62 19.00 25-29 8.20 11.09 58.76 0.19 266.08 972.17 16.54 30-34 8.70 11.76 47.67 0.25 208.96 706.09 14.81 35-39 6.70 9.06 35.91 0.25 156.91 497.13 13.84 40-44 5.45 7.37 26.85 0.27 115.84 340.22 12.67 45-49 3.17 4.28 19.48 0.22 86.71 224.38 11.52 50-54 3.17 4.28 15.20 0.28 65.32 137.67 9.06 55-59 3.17 4.28 10.92 0.39 43.92 72.35 6.62 60-64 3.17 4.28 6.64 0.64 22.52 28.44 4.28 65+ 1.75 2.36 2.36 1.00 5.91 5.91 2.50 74.00 100.00 2815.95

84 APPENDIX 4. LIFE TABLES CORRECTED AFTER RÖSING AND JANKAUSKAS (1997). TABLES 9.1.-9.6. Table 9.1. Rõsna-Saare I barrow cemetery. Corrected life table after Rösing and Jankauskas (1997) x Dx dx lx qx Lx Tx ex 0 27.28 30.00 100.00 0.30 85.00 1867.31 18.67 1-4 13.64 15.00 70.00 0.21 249.98 1782.31 25.46 5-9 6.25 6.87 54.99 0.13 257.78 1532.33 27.86 10-14 3.75 4.12 48.12 0.09 230.28 1274.55 26.49 15-19 5.96 6.56 43.99 0.15 203.58 1044.26 23.74 20-39 15.86 17.45 37.44 0.47 574.25 840.69 22.46 40-59 14.14 15.55 19.99 0.78 244.25 266.43 13.33 60+ 4.03 4.44 4.44 1.00 22.18 22.18 5.00 90.9 1867.31

Table 9.2. Rõsna-Saare II barrow cemetery. Corrected life table after Rösing and Jankauskas (1997) x Dx dx lx qx Lx Tx ex 0 19.45 30.00 100.00 0.30 85.00 1943.76 19.44 1-4 9.73 15.00 70.00 0.21 250.02 1858.76 26.55 5-9 3.96 6.10 55.01 0.11 259.77 1608.74 29.25 10-14 3.71 5.72 48.90 0.12 230.21 1348.97 27.59 15-19 2.94 4.54 43.18 0.11 204.56 1118.77 25.91 20-39 11.78 18.16 38.64 0.47 591.18 914.21 23.66 40-59 9.44 14.57 20.48 0.71 263.89 323.02 15.77 60+ 3.83 5.91 5.91 1.00 29.56 59.13 10.00 64.8 100.00 1914.19

Table 9.3. Suure Rõsna barrow cemetery. Corrected life table after Rösing and Jankauskas (1997) x Dx dx lx qx Lx Tx ex 0 17.78 30.00 100.00 0.30 85.00 2201.11 22.01 1-4 8.89 15.00 70.00 0.21 249.99 2116.11 30.23 5-9 4.79 8.09 54.99 0.15 254.75 1866.13 33.93 10-14 2.79 4.71 46.91 0.10 222.76 1611.37 34.35 15-19 0.50 0.84 42.20 0.02 208.87 1388.62 32.91 20-39 6.17 10.41 41.35 0.25 722.88 1179.75 28.53 40-59 12.50 21.10 30.94 0.68 407.70 456.87 14.77 60+ 5.83 9.83 9.83 1.00 49.17 49.17 5.00 59.2 100.00 2201.11

85 Table 9.4. Rõsna barrow cemeteries (summarised). Corrected life table after Rösing and Jankauskas (1997) x Dx dx lx qx Lx Tx ex 0 64.4 30.0 100.0 0.3 85.0 1975.2 19.75 1-4 32.2 15.0 70.0 0.2 250.2 1890.2 26.99 5-9 15.0 7.0 55.1 0.1 257.8 1640.0 29.79 10-14 10.3 4.8 48.1 0.1 228.4 1382.2 28.76 15-19 9.4 4.4 43.3 0.1 205.5 1153.8 26.65 20-39 33.8 15.7 38.9 0.4 620.9 948.3 24.37 40-59 36.1 16.8 23.2 0.7 295.5 327.4 14.13 60+ 13.7 6.4 6.4 1.0 31.9 31.9 5.00 214.8 100.0 1975.2 Table 9.5. Maidla I stone grave. Corrected life table after Rösing and Jankauskas (1997). x Dx dx lx qx Lx Tx ex 0 9.27 30.00 100.00 0.30 85.00 1991.91 19.92 1-4 4.64 15.00 70.00 0.21 250.02 1906.91 27.24 5-9 4.38 14.16 55.01 0.26 239.65 1656.88 30.12 10-14 0.63 2.02 40.85 0.05 199.20 1417.24 34.69 15-19 0.50 1.62 38.83 0.04 190.10 1218.04 31.37 20-39 2.90 9.37 37.21 0.25 650.50 1027.94 27.62 40-59 6.56 21.24 27.84 0.76 344.42 377.44 13.56 60+ 2.04 6.60 6.60 1.00 33.02 33.02 5.00 30.9 100.00 1991.91 Table 9.6. Maidla II stone grave. Corrected life table after Rösing and Jankauskas (1997) x Dx dx lx qx Lx Tx ex 0 33.36 30.00 100.00 0.30 85.00 1914.94 19.15 1-4 16.68 15.00 70.00 0.21 250.01 1829.94 26.14 5-9 8.17 7.34 55.00 0.13 256.66 1579.92 28.72 10-14 1.00 0.90 47.66 0.02 236.05 1323.27 27.76 15-19 1.81 1.63 46.76 0.03 229.72 1087.21 23.25 20-24 6.70 6.03 45.13 0.13 210.57 857.49 19.00 25-29 8.20 7.38 39.10 0.19 177.06 646.91 16.54 30-34 8.70 7.83 31.72 0.25 139.05 469.85 14.81 35-39 6.70 6.03 23.90 0.25 104.41 330.80 13.84 40-44 5.45 4.90 17.87 0.27 77.08 226.39 12.67 45-49 3.17 2.85 12.96 0.22 57.70 149.31 11.52 50-54 3.17 2.85 10.12 0.28 43.46 91.61 9.06 55-59 3.17 2.85 7.27 0.39 29.23 48.15 6.62 60-64 3.17 2.85 4.42 0.64 14.99 18.92 4.28 65+ 1.75 1.57 1.57 1.00 3.93 3.93 2.50 111.20 100.00 1914.94

86 APPENDIX 5. POANSE DEMOGRAPHIC CALCULATIONS. TABLES 10.1.-10.4

Table 10.1. Poanse I tarand grave. Non-corrected life table. x n dx lx qx Lx Tx ex 0 7.0 15.37 100.00 0.15 92.32 2227.3 22.27 1-4 4.3 9.49 84.63 0.11 319.54 2134.98 25.23 5-9 5.7 12.41 75.14 0.17 344.66 1815.45 24.16 10-14 2.9 6.34 62.73 0.10 297.80 1470.78 23.45 15-19 6.1 13.31 56.39 0.24 248.68 1172.99 20.80 20-24 4.7 10.37 43.08 0.24 189.49 924.30 21.45 25-29 1.2 2.59 32.71 0.08 157.09 734.82 22.46 30-34 1.3 2.92 30.12 0.10 143.33 577.72 19.18 35-39 1.2 2.72 27.21 0.10 129.25 434.39 15.96 40-44 2.8 6.07 24.49 0.25 107.27 305.14 12.46 45-49 2.6 5.68 18.42 0.31 77.89 197.87 10.74 50-54 1.8 3.92 12.74 0.31 53.88 119.98 9.42 55-59 1.3 2.94 8.81 0.33 36.72 66.10 7.50 60-64 1.3 2.94 5.88 0.50 22.03 29.38 5.00 65+ 1.3 2.94 2.94 1.00 7.34 7.34 2.50 46 100 2227.3 Table 10.2. Poanse II tarand grave. Non-corrected life table. x Dx dx lx qx Lx Tx ex 0 3.00 8.57 100.00 0.09 95.71 2511.30 25.11 1-4 6.40 18.29 91.43 0.20 329.14 2415.59 26.42 5-9 2.60 7.43 73.14 0.10 347.14 2086.44 28.53 10-14 3.48 9.94 65.71 0.15 303.71 1739.30 26.47 15-19 2.30 6.57 55.77 0.12 262.43 1435.59 25.74 20-24 3.00 8.57 49.20 0.17 224.57 1173.16 23.84 25-29 0.80 2.29 40.63 0.06 197.43 948.59 23.35 30-34 1.64 4.69 38.34 0.12 179.98 751.16 19.59 35-39 2.09 5.98 33.65 0.18 153.30 571.18 16.97 40-44 2.20 6.28 27.67 0.23 122.65 417.87 15.10 45-49 0.74 2.12 21.39 0.10 101.64 295.22 13.80 50-54 1.68 4.79 19.27 0.25 84.36 193.58 10.05 55-59 1.68 4.79 14.48 0.33 60.44 109.21 7.54 60-64 1.68 4.79 9.69 0.49 36.51 48.78 5.03 65+ 1.68 4.79 4.91 0.97 12.27 12.27 2.50 35.0 2511.3

Table 10.3. Poanse I tarand grave. Corrected lifetable after Rösing and Jankauskas (1997) x n dx lx qx Lx Tx ex 0 18.69 30.00 100.00 0.30 85.00 1664.00 16.64 1-4 9.35 15.00 70.00 0.21 250.01 1579.00 22.56 5-9 5.66 9.08 55.00 0.17 252.31 1328.99 24.16 10-14 2.89 4.64 45.92 0.10 218.00 1076.68 23.45 15-19 6.07 9.74 41.28 0.24 182.05 858.68 20.80 20-24 4.73 7.59 31.54 0.24 138.71 676.63 21.45 25-29 1.18 1.89 23.95 0.08 115.00 537.92 22.46 30-34 1.33 2.13 22.05 0.10 104.93 422.92 19.18 35-39 1.24 1.99 19.92 0.10 94.62 317.99 15.96 40-44 2.77 4.45 17.93 0.25 78.52 223.38 12.46 45-49 2.59 4.16 13.48 0.31 57.02 144.85 10.74 50-54 1.79 2.87 9.33 0.31 39.44 87.83 9.42 55-59 1.34 2.15 6.45 0.33 26.88 48.39 7.50 60-64 1.34 2.15 4.30 0.50 16.13 21.51 5.00 65+ 1.34 2.15 2.15 1.00 5.38 5.38 2.50 62.31 100.00 1664.00 Table 10.4. Poanse II tarand grave. Corrected life table after Rösing and Jankauskas (1997) x Dx dx lx qx Lx Tx ex 0 13.94 30.00 100.00 0.30 85.00 1900.77 19.01 1-4 6.97 15.00 70.00 0.21 250.00 1815.77 25.94 5-9 2.60 5.60 55.00 0.10 261.01 1565.77 28.47 10-14 3.48 7.49 49.40 0.15 228.30 1304.76 26.41 15-19 2.30 4.95 41.92 0.12 197.20 1076.46 25.68 20-24 3.00 6.46 36.97 0.17 168.69 879.25 23.79 25-29 0.80 1.72 30.51 0.06 148.24 710.56 23.29 30-34 1.64 3.53 28.79 0.12 135.10 562.32 19.53 35-39 2.09 4.50 25.25 0.18 115.01 427.21 16.92 40-44 2.20 4.73 20.75 0.23 91.92 312.21 15.05 45-49 0.74 1.60 16.02 0.10 76.09 220.28 13.75 50-54 1.68 3.60 14.42 0.25 63.08 144.19 10.00 55-59 1.68 3.60 10.81 0.33 45.06 81.11 7.50 60-64 1.68 3.60 7.21 0.50 27.04 36.05 5.00 65+ 1.68 3.60 3.60 1.00 9.01 9.01 2.50 46.5 1900.8

88 PUBLICATIONS

I Allmäe, R. (2010). Some remarks on Kaseküla stone-cist grave, Läänemaa, Estonia. Fennoscandia Archaeologica, XXVII, 45–52. Fennoscandia archaeologica XXVII (2010)

Raili Allmäe SOME REMARKS ON KASEKÜLA STONE-CIST GRAVE, LÄÄNEMAA, ESTONIA.

Abstract The stone-cist grave of Kaseküla, Läänemaa, Western Estonia, is an intriguing archaeological object regarding human remains. The number of infant burials in the grave is remarkable. The results of osteological analysis and radiocarbon dating (AMS) of human remains of this site are presented in this paper. The phenomenon of numerous infant burials and (re)usage of ancient burial places will be discussed, also what care should be taken regarding the demographic estimations and models based on human remains of ancient burial places.

Keywords: stone-graves, secondary burials, infant burials, reuse of graves

Raili Allmäe, Institute of History, Tallinn University, Rüütli 6, 10130 Tallinn, Estonia. E-mail: [email protected]

INTRODUCTION iron shepherd’s crook pins are the latest fi nds in stone-cist graves according to current data (Lang Stone-cist graves are above-ground structures that 2007b: 161). Mandel (1975) dated the burial site have one or several cists in the middle and which to the second half of the fi rst millennium BC (Pre- are enclosed by one or several circular stone walls, Roman Iron Age); later on, Lang (1996: 297) dated fi lled with soil and stones and covered with a stone it to the Late Bronze Age (1100–500 BC) and also heap (Lang 2007b: 147). Stone-cist graves appear suggested the possibility of secondary burials in the into Estonian landscape with the transition to the grave. It is not surprising that only a few artefacts Late Bronze Age (1100–500 BC) and are more were found in the Kaseküla stone-cist grave. The characteristic to the northern and western coastal Estonian stone-cist graves do not contain many areas of Estonia (Lang 2007b). Archaeologist grave goods; approximately one in three graves has Mati Mandel investigated the Kaseküla stone-cist a cist(s) that contains fi nds that can be identifi ed grave area in western Estonia in 1973 (Figs. 1–2). and dated, whereas fi nds are twice as frequent in Inside the central stone-cist commingled unburnt the burials outside the cist (Lang 2007b: 155). human bones and a fragment of bronze razor were The osteological material from the site was found; outside (north) of the stone encirclement analysed by Jonathan Kalman at the end the the remains of another human burial were found 1990s. The central cist contained the remains of (Mandel 1973; Mandel 1975: 76). The most nu- a man aged about 50 years and the remains of at merous artefacts between the central cist and the least three infants; within the stone encirclement circular wall were pieces of pottery, also some burials of at least 16 infants and one 3–4 year old pieces of iron and a fragment of a iron shepherd’s child were found; outside (north) of the stone en- crook pin were found (Mandel 1975: 76). Unfor- circlement the remains of at least two more adults tunately the archaeologists have not yet dated the lay, although the majority of at least one skeleton ceramics found in Kaseküla stone-cist grave, thus was absent (Kalman 2000). In his paper, Kalman its connection to primary or secondary burials is (2000) has proposed three explanations why such not ascertained. The form of shepherd’s crook pins a remarkable number of infants have been buried reached Estonia from Ukraine and Belarus during there: a special burial place for children, human the middle Pre-Roman Age (Lang 2007b: 184); sacrifi ce or epidemics.

93 the deposited material for radiocarbon dating in summer 2009: a fragment of the cranial vault of an adult from the central stone-cist (fi nd No 29), a left femoral bone of an infant in the central stone- cist (fi nd No 29) and the petrous part of the left temporal bone of an infant from the encirclement area (fi nd No 40). The author analysed the bone material in summer 2009. Because a discrepancy occurred between the former (Kalman 2000) and the new results of the analysis, the material was analysed once again in autumn 2009. A lot of bone samples were taken by another researcher between these two osteological analysis in 2009. On the one hand, the commingled osteological material is quite poorly preserved – the bones are fragmented and eroded, on the other hand, the remains of infants are numerous and some Fig. 1. Location of Kaseküla stone-cist grave. bones even measurable. The ossifi cated thyroid cartilage found in the stone-cist also refers to The author of the present paper has studied a good excavation technique and preservation cremated osteological materials of several western of osteological material. In the 1970s the soil Estonian stone-graves of the Middle and Late Iron removed from archaeological objects was not Age (Allmäe 2003) and has made some demo- sieved before disposal, this explains why most graphic estimations on the basis of the osteologi- of the (expected) human teeth are absent amongst cal materials of stone-grave 2 of Maidla (Allmäe the deposited material. 2006). It was noticed that the cremated remains The method of recurrent bone fragments was of infants and children amongst analysed crema- used to determine the minimum number of burials. tions were scant. The results of the osteological The sex and age of the deceased was determined study of the Kaseküla stone-cist grave (Kalman according to common osteological standards 2000) were intriguing and raised a question: do (Ubelaker 1978; WEA 1980; Brothwell 1981; the remains of infants buried there date from the Buikstra & Ubelaker 1994; Bass 2005; Mays Bronze Age or from later periods? Unfortunately, 2006). The maximum lengths of the diaphyses of the placement of infant burials in the stone-cist long bones of infants were measured. The values grave area was not published by Kalman (2000), were compared to the standard correlation be- which made it necessary to analyse the osteologi- tween age estimates and maximum lengths of long cal material once again. bone diaphyses of infants (Lovejoy et al. 1990) The present study offers new analysis of the hu- and to the same correlations compiled on the basis man remains of the Kaseküla stone-cist grave, the of Italian and Estonian archaeological materials radiocarbon dating of bone material is presented, (Allmäe 1998; Facchini & Vesci 2004). and some issues concerning the use of stone- Three samples of human bones were radiocar- graves, burials of infants and the demographic bon dated in Poznan Radiocarbon Laboratory by estimations based on ancient burial places are AMS method (calibration was made by OxCaL discussed. 3.10 software, Bronk Ramsey 2005, atmospheric data from Reimer et al. 2004). OUTLINE OF MATERIAL AND METHODS RESULTS OF OSTEOLOGICAL ANALYSIS The fragmentary and commingled bone material from the Kaseküla stone-cist grave area (Fig. The minimum number of individuals (28) was 2) has been gathered by squares (2 x 2m). The determined on the basis of petrous portions’ of number of the collected bone fi nds is 81, and the the temporal bones in the osteological material material is deposited in the Estonian History Mu- collected; of which the vast majority (25, right seum. Three pieces of bones were removed from side) belonged to infants. Four infants were found

94 in the central stone cist and 21 inside the stone stone-cist. Kalman (2000: 20) distinguished the encirclement of the grave (Fig. 2). As expected, other skeleton outside of the stone encirclement other skeletal parts of infants – fragments of long on the basis of one piece of cranium – the frag- bones, a cranial vault, facial bones, mandibles, ment from the nuchal area. Thus, the new analyses ribs, vertebral bodies and arches, and so forth – indicated the presence of two male skeletons in were found in the stone cist and, also, in larger the grave area too, but the presence of the bones quantities elsewhere in the stone encirclement of a third adult was not conﬁ rmed. Amongst the area. The diaphyses of long bones (10) were bone material three petrous parts (right side) of measured to certify that the remains belong to an the temporal bones were found, which may belong infant age group. The maximum lengths never to adults or even to children. exceeded the variation range characteristic to According to the results of the new osteologi- infants under 6 months old according to Lovejoy cal analysis the remains of two adults have been et al. (1990), but were in some cases below the distinguished: the fragments of bones and teeth variation range. At the same time these values of a male skeleton were found in the north of the were in the range of infants under 6 months old stone encirclement area (squares 4–5/p) where of Italian (Bologna) Modern sample and Esto- the remains of the second cist were found (Fig. nian archaeological Medieval and Modern series 2, A). The remains of an adult male were found (Allmäe 1998; Facchini & Vesci 2004). The few in the central cist and its surroundings, in squares available measurements of long bones suggest that 8–9–10/o–q–p (Fig. 2, B). Both men were of old perinatal deaths were more plausible for infants age, 50 years and older. The burial of a 3–4 year buried in the Kaseküla old child was conﬁ rmed, the skeletal parts and Kalman (2000) has suggested that the remains teeth were found in square 9/q (Fig. 2, C). It de- of three adults have been buried on the site: an serves to be noted that the teeth of the buried male adult male in the central cist and two adults outside B were also found outside of the cist, in squares (north) of the stone encirclement, where Mandel 10/n and 10/o, probably caused by later digging (1975: 76) found the remains of another probable into the cist. The teeth of two other skeletons, A

k l m n o p q r

4 A

X X X X X 8 X X X X X X X X B X X X X 9 XX X C X X X X

X 10 Fig. 2. The distribution of burials in the Kaseküla stone-cist grave

11 area (adapted from Mandel 1973). A: Male aged 50 years or more, B: male aged 50 years or more, C: 3–4

12 1 m year old child, X: infants, location in square is ﬁ ctional.

95 Table 1. Radiocarbon (AMS) dates of Kaseküla bone samples. Lab. No. Uncal. BP Calibrated 1/2 Context Material N/C (%) Poz-32412 2780±30 980–895/1010–840 BC Stone cist, find #29 Cranial vault, adult 1.7/6.7 Poz-32413 1195±30 760–900/710–940 AD Stone cist, find #29 Left femur, infant 3.0/9.2 Poz-32414 920±30 1040–1160/1020–1190 AD Square 9/q, find #40 Left temporal bone1, infant 3.2/11.3 1) Petrous part (squares 4–5/p) and C (square 9/q), were probably found inside the Late Bronze Age stone-cist lying in the area, where the body decomposed (Table 1; Fig. 3). Lang (2007b:154) has pointed (Fig 2). In short, 28 burials we estimated from out that the number of burials in Estonian stone- the grave area, including 25 infants, one child cist graves varies from several to tens of burials aged 3–4 years and two male adults older than and is in good correlation with the duration of the 50 years. In the central stone cist the remains of use of the graves; the graves often contain grave four infants were buried. The remains of at least goods from the Late Bronze Age to Roman Iron 21 infants were found around the central cist in Age. The radiocarbon dating of human bones the stone encirclement area. It is noteworthy that proved that in stone-cist graves the burials may all infant burials were more or less related to the originate from later periods of the Iron Age, in central cist (Fig. 2). the case of the Kaseküla stone-cist grave from Late Iron Age. RADIOCARBON DATING OF INHUMATED HUMAN REMAINS DISCUSSION

The remarkable number of buried infants in the Often observations and arguments are presented Kaseküla stone-cist grave was intriguing. The ex- in literature that in some burial places the number amination of the Kaseküla bone material in sum- of infants and children is underrepresented. The mer 2009 proved that the skeletal parts of infants proposed reasons to explain this phenomenon vary were somewhat better preserved in comparison from suggesting that the bones of infants (burnt with the adult burial. This intrigued to date the and unburnt) were too fragile to survive (Walker et adult burial in the central cist and also some infant al. 1988; Holck 1995), the infants may have been bones in grave. Are the remains of infants buried buried elsewhere, or have been lost in cemeter- at the same time with the adult in the cist? ies due to continuous use of cemeteries or false Two bone samples were taken from the stone- excavation techniques (Molleson 1991; Saunders cist; the third sample was taken from the stone 1992; Roberts & Manchester 1995; Mays 2006). encirclement area (Table 1). The results of ra- Stig Welinder (1998: 187) argues that there is no diocarbon analysis date the adult burial in the notable lack of children’s graves, when all kinds stone-cist to the Late Bronze Age, as suggested of burials are included in the analysis. earlier by Lang (1996: 297). Other results of Secondary burials in prehistoric stone graves radiocarbon dating were surprising; the infants are known in Estonian archaeology (Jaanits et al. have been buried to the grave area approximately 1982: 150, 178; Lang 2000: 147; Mandel 2003: 2000 years later, including the remains of infants 140; Lang 2007a: 57– 9, 191–2; Vedru 2009), but

Fig. 3. Calibrated dates of Kaseküla bone samples.

96 the subject has not been studied comprehensively. of child inhumations was rather representative Recently Tõnno Jonuks (2009: 175) has discussed (Allmäe 2003; 2006). The author has also studied the absence of research concerning secondary cremations of the Culture of Long Barrows from burials in Estonian prehistoric burial grounds. the second half of the fi rst millennium in northern Jonuks (2009) pointed out that the importance of Setumaa (south-east Estonia). There the cremains once established burial places for our ancestors of infants and children were commonly buried can be connected especially to secondary burials into the family or community graves (Allmäe & in the graves. There have been discussions (Parker Maldre 2005; Allmäe et al. 2007, Allmäe et al. Pearson 1999: 17) that funerary archaeologists 2008; Aun et al. 2008). The differences in west- concentrate on the period when the cemetery was ern and south-eastern Estonian cremations are in use, but give less attention to why it fell out of obvious and it leads to another research problem: use or was abandoned, or re-used after some time are the remains of infants and children of Late again – not suffi cient attention is paid to important Iron Age in western Estonia treated differently aspects of cemetery dynamics. and/or were buried somewhere else than their The reuse of the Kaseküla stone-cist grave family grave? Is it possible that the inhumations nearly 2000 years later has been proved by ra- of infants in stone-grave 2 of Maidla are from a diocarbon dating. The phenomenon to use ear- later period? lier stone graves is well-known from Sweden Cremation burials into grave 2 in Maidla began (Welinder 1998: 188–9), the similarity to the during the 10th century (Mandel 2003: 59). Mandel Kaseküla grave can for example be observed in has suggested (2003:176) that the inhumations the Bronze Age cairns in Gotland, where tens in grave 2 in Maidla could have been later ones, of infants have been buried into the cairns, the probably from the end of 12th century or from the radiocarbon dating suggest that the cairns were fi rst half of 13th century. The radiocarbon dating used during the Middle Ages as an alternative to of one infant inhumation from grave 2 in Maidla Christian churchyards (Lindqvist 1981). (unpublished data of author) referred to an even The grave of Kaseküla had probably acquired later period, thus secondary burials in the grave a different signifi cance after 2000 years, but we are likely. The occurrence of secondary infant do not know the reason why – unfortunately we burials in grave 2 of Maidla demonstrated the can only assume what our ancestors thought at same pattern - later reuse of the grave, of course the time. The stone-cists graves are visible in the time-span between primary and secondary contemporary landscape as round heaps covered burials in Kaseküla stone-cist grave is remarkably with sod that are approximately 10–15 m in di- longer. The secondary infant burials in ancient ameter and up to 1,5 m in height (Lang 2007b: graves have not been previously distinguished in 148). The signifi cance of monumental graves and Estonian archaeology with radiocarbon dating the landscapes for our ancestors has been lately of bones. Radiocarbon dating (AMS) of human discussed by Gurly Vedru (2009), she suggested bones (cremated as well un-cremated bones) is a that the later reuse of burial sites was due to the good method to distinguish later and/or second- need to manifest oneself through ancestors and/ ary burials in graves, especially when there is or place. It is possible that the act of burial also no, few or confusing archaeological artefacts in served to physically ‘plant’ the dead into land, graves. Unfortunately the radiocarbon dating of making the remains an inalienable and fi xed part inhumated infant burials was not available when of that land (Parker Pearson 1999: 17). The use the demographic estimations were compiled for of megaliths as burial grounds for infants, and for stone-grave 2 of Maidla (Allmäe 2006). other ambiguous categories of individuals, leads The use of old graves for burying the remains to re-think both the meaning and signifi cance of of infants and small children is not an uncommon the original monument (Finley 2000: 420). The habit in human societies. Cillinis, special burial infant remains in the Kaseküla stone-cist grave grounds for children are for example well-known originate from the Late Iron Age, when the pre- from Early Christian Ireland, but also the dif- vailing mortuary custom in western Estonia was ferential treatment of children’s remains in the cremation. For example, the number of child Neolithic and Bronze Age is known – the reuse of cremations was few in stone-grave 2 of Maidla earlier monumental graves like Megalithic tombs from the Late Iron Age, whereas the number to bury infants was practised (Finley 2000). In

97 Christian societies, un-baptised people, suicide strated in the Lastekangrud stone-cist graves in victims and other people who violated Christian Rebala, northern Estonia. Vello Lõugas (1983) practises were not buried into consecrated land archaeologically excavated the stone-graves of around churches and chapels (Purhonen 1998: Rebala Lastekangrud in 1982. Osteological analy- 119–20). This brought along the necessity to ses determined a comparatively high number separate burial grounds and different funerary of infant burials – 17 inhumations of children practices for these members of the society. The and juveniles/young adults amongst 24 burials reasoning why infants were not buried into regular (Kalman 1999). The stone-cist graves were and/or family graves in prehistoric times is com- archaeologically re-investigated in 2000 (Lang plicated, as we do not discern the beliefs, rules et al. 2001). The recent result of osteological and attitudes of ancient people. analyses is even more surprising; the total number Taking into account that remains of infants and of inhumation burials in graves was 40, and of children were not cremated and not buried into them 23 were children’s burials (up to 14 years) family graves as a rule, but rather somewhere and 7 juveniles/young adults (up to 22 years); the else, the Kaseküla stone-cist grave area might be number of infants increased from 6 to 15 (Kalman an alternative burial place for infants of the Late 1999; Lang et al. 2001: 39–45). Once again, the Iron Age community. It is noteworthy that next plausible later burials in the area of stone-cist to the Kaseküla stone-cist grave there is a Late graves and possible segregation of burial customs Iron Age stone-grave (No 14) with cremation according to the age of deceased offer intriguing burials (Mandel 1975: 76; Mandel 2003: 108), further studies. not totally excavated and not yet osteologically It is vital to continue research of such burial investigated. places with noteworthy bias in age structure, Per Holck (1995) has proposed that the crema- especially of these burial grounds where the tion of newborns was avoided, because it was a number of infants and children is over- or under- time and resource consuming process. Were the represented. The radiocarbon dating of human remains of infants buried elsewhere to avoid too remains is a good method to understand the usage much trouble with cremation? This is one possi- and reusage of prehistoric burial grounds. The bility. Was the Kaseküla stone-cist grave a secret results of research may provide us important new place to bury still-born and unexpectedly died data regarding ancient burial customs, and may infants or to hide the consequences of infanticide help us to understand our ancestors’ beliefs and during hundreds of years? The concepts why attitudes towards life and death. infants have been treated differentially and were buried into unusual places in prehistoric times CONCLUSIONS has been discussed by many authors; the ideas vary from human sacrifi ce to hiding the remains The new analysis of osteological material and of unwanted children – victims of infanticide radiocarbon dating of human bones confi rmed (Green 1998; Purhonen 1999; Parker Pearson the existence of Late Iron Age infant burials in 1999; Mays 2006; Chamberlain 2006; Faerman the Kaseküla stone-cist grave of the Late Bronze & Smith 2008). Several causes and reasoning Age. It also provided new important information for infanticide have been suggested, including concerning the Late Iron Age burial customs in adaptive behaviour under certain conditions, like western Estonia. The acquired knowledge also the need to control fertility and reproduction, reminds us to stay critical towards material we control resources, eliminate disabled offspring, work with. It is essential to consider all the facts manipulate sex ratios, eliminate possible illegiti- carefully before making palaeo-demographic estimate offspring and ritual sacrifi ce (Chamberlain mations and reconstructing lifecycles on the basis 2006: 171). of single graves of our ancestors. The radiocarbon Stig Welinder (1998: 189–90) has studied the dating of human bones is a good tool to date graves selection of mortuary practices of children and when there are none, few or confusing artefacts in is suggesting that the burial ritual presumably the grave. It also helps to specify probable later corresponds to steps in the growing-up process burials, when there is a bias in the age structure of children. Again, quite a good example of such of skeletal sample. The mortuary practices and biased age structure of skeletal sample is demon- lifecycles of our ancestral communities is a puzzle

98 that we probably piece together for a long time. Jener, S. Muriel & C. Olaria (eds.), Nasciturus The research would be more fruitful if based on puerulus vobis mater terra: la ueret en la infancia. Diputació de Castelló: Servei d’Investigacions a certain region, comprised different burial places Arqueològiques i Prehistòriques: 211–30. Servei and were fairly retrospective. The regional (bio) d’Investigacions Arqueològiques i Prehistòriques archaeological studies then enable us to follow the (SIAP ), Diputació de Castelló. http://dialnet.uniri- transitions of burial customs, the possible reuse oja.es/servlet/oaiart?codigo=2796721 Finley, N. 2000. Outside of life: tradition of infant burial of burial grounds and to get more reliable data for in Ireland from Cillin to Cist. World of Archaeology palaeo-demographic estimations. 31(3): 407–22. Green, M. 1998. Humans as ritual victims in the later pre-history of Western Europe. Oxford Journal of ACKNOWLEDGMENTS Archaeology 17(2): 169–89. Holck, P. 1995. Why are small children so seldom found The author is very grateful to archaeologists Mati in cremations? In E. Smits, E. Iregren & A.G. Drusini Mandel and Jüri Peets for encouraging me to write (eds.), Cremation Studies in Archaeology: Proceed- ings of the Symposium Amsterdam 26.–27. October this paper. The author is also thankful to Jüri Peets 1995: 33–8. Logos Edizioni, Padova. and reviewer for constructive critics, to Reet Maldre Jaanits, L., Laul, S., Lõugas, V., Tõnisson, E. 1982. and Jaana Ratas for preparing the illustrations and to Eesti esiajalugu. Eesti Raamat, Tallinn. Jonuks, T. 2009. Eesti muinasusund. Dissertationes Ar- Helle Solnask for revising English. The study was chaeologiae Universitas Tartuensis 2. Tartu Ülikooli undertaken in the frame of target funded research kirjastus, Tartu. project of Estonian Government (SF0130012s08) Kalman, J. 1999. Human remains from the stone-cist and supported by the research grant from Estonian graves of Rebala Lastekangrud, North Estonia. Jour- nal of Estonian Archaeology 3(1): 19–34. Science Foundation (N 6899). Kalman, J. 2000. Skeletal analysis of the graves of Kaseküla, Poanse I and Poanse II. Eesti Ajaloo REFERENCES Muuseum: töid ajaloo alalt 2: 17-40. Lang, V. 1996. Muistne Rävala: muistised, kronoloogia ja maaviljelusliku asustuse kujunemine Loo- de-Eestis, eriti Pirita jõe alamjooksu piirkonnas. Unpublished sources Muinasaja Teadus 4(I–II). Teaduste akadeemia Mandel, M. 1973. Aruanne Kaseküla kivikirstkalme kirjastus, Tallinn. kaevamistest 1973. aastal. Eesti Ajaloo Muuseumi Lang, V. 2000. Keskusest ääremaaks: viljelusmajandus- Arhiiv, Tallinn. liku asustuse kujunemine ja areng Vihasoo-Palmse piirkonnas Virumaal. Muinasaja teadus 7. Ajaloo instituut, Tallinn. Lang, V., Laneman, M., Ilves, K. & Kalman, J. 2001. Literature Fossil ﬁ elds and stone-cist graves of Rebala revis- Allmäe, R. 2003. Läänemaa 5.–13. sajandi kalmete an- ited. Archaeological Field Works in Estonia 2000: tropoloogiline aines. In M. Mandel (ed.), Läänemaa 34–47. 5.–13. sajandi kalmed: 243–62. Eesti Ajaloomuu- Lang, V. 2007a. Baltimaade pronksi- ja rauaaeg. Tartu seum: töid ajaloo alalt 5. Eesti ajaloomuuseum, Ülikooli Kirjastus, Tartu. Tallinn. Lang, V. 2007b. The Bronze Age and Early Iron Age Allmäe, R. 2006. Grave 2 of Maidla: the burial site of in Estonia. Estonian Archaelogy 3. Tartu University a single family. Estonian Journal of Archaeology Press, Tartu. 10(1): 3–23. Lindquist, M. 1981. Mylingar offer: utsatta barn eller Bass, W. 2005. Human Osteology: A Laboratory and förhistorika barnegravingsar? Gotländsk arkiv 53: Field Manual. Missouri Archaeological Society, 7–12. Missouri. Lovejoy, C.O., Russell, K.F. & Harrison, M.L. 1990. Brothwell, D.R. 1981. Digging up Bones. British Mu- Long bone growth velocity in Libben Population. seum (Natural History), London & Oxford. American Journal of Human Biology 2: 533–41. Buikstra, J.E. & Ubelaker, D. (eds.) 1994. Standards Lõugas, V. 1983. Über die Steingräbergruppe Laste- for Data Collection from Human Skeletal Remains: kangrud Rebala. ENSV TA Toimetised, Ühiskon- Proceedings of a Seminar at the Field Museum of nateadused 4: 295–7. Natural History. Arkansas Archaeological Survey Mandel, M. 1975. Ausgrabungen Der Steingräber von Press, Fayetteville. Kaseküla. ENSV TA Toimetised, Ühiskonnateadused Chamberlain, A. 2006. Demography in Archaeology. 4: 74–6. Cambridge University Press, New York. Mandel, M. 2003. 5.-13. sajandi Läänemaa kalmed. Facchini, F. & Veschi, S. 2004. Age determination Eesti Ajaloomuuseum: töid ajaloo alalt 5: 1–243. on long bones in a skeletal subadults sample (b- Mays, S. 2006. The Archaeology of Human Bones. 12years). Collegium Antropologicum 28(1): 89–98. Routledge, London & New York. Faerman, M. & Smith, P. 2008. Has society changed Molleson, T. 1991. Demographic implications of the its attitude to infants and children? Evidence from age structure of early English cemetery samples. archaeological sites in the southern Levant. In F.G. Actes des Jornes Anthropologiques 5: 113–21.

99 Parker Pearson, M. 1999. The Archaeology of Death Ubelaker, D.H. 1978. Human Skeletal Remains: Analy- and Burial. Sutton Publishing, Stroud. ses, Interpretations. Aldine, Chicago. Purhonen, P.1998. Kristinukon saapumisesta Suomeen: Vedru, G. 2009. Layers of landscape, layers of site. uskontoarkeologinen tutkimus. Suomen Muinas- Estonian Journal of Archaeology 13(1): 21–35. muistoyhdistyksen Aikakauskirja 106. Suomen Walker, P.L., Johnson, J.R. & Lambert, P.M. 1988. Age muinasmuistoyhdistys, Helsinki. and sex biases in the preservation of human skeletal Purhonen, P.1999. Suku käski suolle viedä. Tieteessä remains. American Journal of Physical Anthropol- tapahtuu 3/1999: 17–20. http://www.tieteessatapah- ogy 76: 183–8. tuu.ﬁ /993/purhonen.htm WEA 1980 = Recommendations 1980. Workshop of Roberts, C. & Manchester, K. 1995. The Archaeology of European anthropologists, ‘Recommendations for Disease, 2nd edition. Sutton Publishing, Stroud. Age and Sex Diagnoses of Skeletons’. Journal of Saunders, S.R. 1992. Subadult skeletons and growth Human Evolution 9: 517–49. related studies. In S.R. Saunders & A. Katzenberg Welinder, S. 1998. The cultural construction of child- (eds.), Skeletal Biology of Past Peoples: Research hood in Scandinavia, 3500 BC–1350 AD. Current Methods: 1–20. Wiley-Liss, New York. Swedish Archaeology 6: 185–204.

100 II Valk, H., Allmäe, R. (2010). Siksälä Kirikumägi: evidence of a new grave form of South-Eastern Estonia? Estonian Journal of Archaeology, 14(1), 40–55.          

    

           

                                                                            

                              

   

                                                                                                                                                                                                                          

103    

          

                           

104      

                                                                                                                                                                                                                                                                                                                          

   

                                                                                                                                                                                                                                                      

105    

                                                                                                                     

  

                                                                                

            

                                                                                                           

             

106      

                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                

107    

                                                                                                                                                                                                                                                                                                                                                    

                  

                                                                                           

108      

                                                                                                                                                                                                                                                                                                                                                                              

 

                                 ±                                                                    ±                 

109    

                                              

  

                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                

110      

                                     

                                                                                                                                                                                                

111    

                                                                                                                                                                                                                                                                        

     

                                                                                                                                                                                                                                     

      

112      

                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               

   плщдк        

113    

                                                                                                                                    



                                                                                                                                                                                                                                                                                  



                                                      

114      



                                                                                                         ллмяэ Р Аун М  Малдре Л Предварительные результаты изучения остеолгического материала курганных могильников РыснаСааре  и  в Северной Сетумаа ЮгоВосточная Эстония  Археология и история Пскова и Псковской земли Семинар имени академика В В Седова Материалы  заседания посвященного памяти профессора А Р Артемьева Институт Археологии Российской Академии Наук Псковский Археологический Центр Псков      Аун М Археологические памятники второй половины го тысячелетия нэ в ЮгоВосточной Эстонии Олион Таллинн                                                                                                                                                                                                                                                                                             Лопатин Н В Проблема соотношения Пскова и Изборска в  вв                                                            

115    

                        Рябинин  А Жальничные могилы Ижорской возвышенности  Новое в археологии Северозапада СССР  В М Массон Наука Ленинград       Рябинин  А Водская земля Великого Новгорода результаты археологических исследований  гг Дмитрий Буланин СПетербург                                                                                                                                                                                                           

    

    



                                                                                                                                                               

116      

                                                                                                                                                                     ±                                                      ±     ±                                                                                                                                                                                                                               

117    

                                                                                                                                                                                                                                                                                                                                                                                                            

118 III Allmäe, R. (2013). Observations on Estonian Iron Age cremations. Archaeologia Baltica, 19, 31–47. OBSERVATIOnS On ESTOnIAn IROn AGE CREMATIOnS

RAILI ALLMÄE BALTICA 19 BALTICA

Abstract

Several Estonian burial places with cremations were investigated in the period 1997 to 2011. during the research, various descriptive and metric data on cremated bone materials was observed. The present paper is an attempt to systematise and interpret the data collected, in order to provide some generalisations on Estonian cremations. A comparative study of graves on the basis of the minimum number of buried individuals and the number of determined bone finds in graves, as well as ARCHAEOLOGIA bone fragmentation, is presented. Radiocarbon dating (AMS method) of burnt human bones from six investigated graves was conducted in order to specify the usage time of the graves. Some conclusions on possible temporal changes and cultural differences in burial practices are made on the basis of these characteristics. Key words: cremation, Iron Age, bone fragment size, radiocarbon dates, cremated bone, Estonia.

SOCIETIES Introduction anthropological analyses was to estimate the number Of THE PAST: of buried people, their age and sex. In addition, two APPROACHES TO LAndSCAPE Osteological material from Iron Age cremations in Es- conclusions were made on the basis of osteological tonia has been studied since the 1990s. The analyses analyses. first, the cremation temperature was high: of cremated bones are mostly macroscopic; the main the bones were all white, calcinated and highly frag- aim of the analyses has been the identification of the mented. The second conclusion was that the majority of number of buried people, their age at death, and their the identified bones belonged to animals; human bones sex. The anthropological identifications are then com- were rarely identified in the graves. In 1998, coopera- bined with archaeological data to interpret the artefacts tion with the archaeologist Mati Mandel started; Man- in the graves and the burial customs of the period and/ del has excavated west Estonian cremations since the or cultural area. Less attention has been paid to the col- 1970s (Mandel 2003). The cremated and non-cremated our and fragmentation of the bones, overall descrip- human and animal bones of eight fifth to 13th-century tions of deposited bone material, burial practices, and stone graves were analysed, including the analysis of palaeodemographic analyses. long-bone fragment size (Allmäe 2003; Maldre 2003). The first known study of cremated human bones in Est- In his study of the fifth to 13th-century stone graves of onia was conducted by Ken Kalling (1993), who ana- Läänemaa (western Estonia), Mandel (2003) combined lysed cremated human remains (cremains) from Viimsi the archaeological and osteological data. In 2006, a de- I tarand grave, dated to 350–500 Ad (Lang 1993, p.55; mographic analysis based on anthropological materials 2007, p.133). Anu Kivirüüt (2011) has revisited this for the Maidla II stone grave was published (Allmäe material. 2006). Since 2004, cremains from Long Barrow Cul- ture (fifth to tenth centuries) sand barrows in southeast Jonathan Kalman (2000a; 2000b; 2000c) has made Estonia have been analysed. Mare Aun excavated the determinations of cremated bones which occurred in sand barrow cemeteries between 1970 and 1980 (Аun Early Iron Age tarand graves. Leiu Heapost (2007) 1992). The combined study of the results of the analy- identified cremations at Kalmetemägi, Siksälä, in ses of burial archaeology and osteology are published southeast Estonia. in several papers, including an analysis of bone frag- In 1998, Marika Mägi excavated the Viking Age graves ment sizes (Allmäe, Maldre 2005; Allmäe et al. 2007; of Piila. The cremated bone material was analysed by Aun et al. 2008; Allmäe et al. 2009). the archaeozoologist Liina Maldre and the author of In 2008, the cremains of the Uugla III stone grave the present paper (Mägi et al. 1998). The idea of the (11th to 13th centuries) in western Estonia were analysed and published (Mandel, Allmäe 2009). In 2009,

121 the cremains of the Kirikumägi flat ground cemetery 268 finds, which were collected in two-by-two-metre (990–1160 Ad) in southeast Estonia were analysed square plots. The grave generally contained scattered (Valk, Allmäe 2009). In 2010, the archaeological and cremation burials, although some inhumations were anthropological study of Kirikumägi flat ground ceme- also found (Mandel 2003; Allmäe 2003). tery was published, including an analysis of the colour The second stone grave of Maidla was used mainly and size of bone fragments (Valk, Allmäe 2010). from the tenth century to the beginning of the 13th during the 1997 to 2011 research period, various de- century Ad (Mandel 2003, p.59). The bone material scriptive and metric data on cremated bones was col- consists of 674 finds, which were collected in two-by- lected. The present paper is an attempt to analyse and two-metre square plots. The grave contained inhuma- Observations on Estonian Iron Observations on Estonian Iron Age Cremations systematise this data, and to make some generalising tions, as well as scattered cremation burials (Mandel observations on Estonian burial places with crema- 2003; Allmäe 2003). tions. The radiocarbon dates (AMS method) of burnt • Kirbla stone grave, dating from the 11th or 12th

RAILI ALLMÄE human bones from the six investigated graves are also century Ad, was excavated by Mati Mandel in presented. Some conclusions on temporal changes 1983 (Mandel 2003, p.90). A total of 70 bone finds in cremation burial practices are made based on cor- were collected in two-by-two-metre square plots. relations between the dates and the bone fragmenta- The human bone material consisted only of burnt tion stage in different graves. A comparative study bones (Mandel 2003; Allmäe 2003). of the graves on the basis of other characteristics, for example, the minimum number of buried individuals, • Uugla III stone grave, dating from the 11th or 12th the number of determined bone finds among the total century Ad, was excavated by Mati Mandel in number of bone finds, and the colour of the cremated 2008. A total of 326 bone finds were collected in bone fragments in graves, is also presented. The colour two-by-two-metre square plots. The material con- of the bones has a descriptive value in this study, and is sisted of only cremated bones (Mandel, Allmäe only briefly discussed. 2009). The stone graves analysed in western Estonia were Material all of an irregular stone construction (Mandel 2003, p.128, pp.132-133). The list of analysed material from western Estonia is The list of analysed material from southeast Estonia is as follows (fig. 1): as follows (fig. 1): • Ehmja stone grave, from the fifth to seventh cen- The Põlgaste tarand grave, dating from the third to the turies Ad and from the tenth to 13th centuries Ad, fifth centuries AD (Early Iron Age), was excavated were excavated by M. Mandel between 1982 and by Silvia Laul between 1970 and 1973 (Laul 2001, 1991. A total of 248 bone finds were collected p.40ff). Tarand graves are burial places with charac- from square plots two metres by two metres. Eight teristic quadrangular stone enclosures, commonly with burial pits were recorded, probably from the fifth a north-south orientation (Jaanits et al. 1982, p.207; to the seventh centuries. In addition, scattered cre- Lang 2007, pp.170, 192). The bone material of Põl- mation burials from the tenth to the 13th centuries gaste grave was collected by plots of one square metre, were recorded from the upper layers of the grave. and is all cremated (Laul 2001, p.42). The cremains of Ehmja stone grave was repeatedly disturbed dur- Põlgaste tarand grave were analysed in 2010 and 2011. ing Prehistoric times. Mandel (2003) has suggest- Unfortunately, the author could not find any whole ed it happened during the reuse of the cemetery. bone material excavated in the grave. The only bone The grave contained cremated and uncremated hu- finds available were collected from the western, earlier man bones (Mandel 2003; Allmäe 2003). part of the grave, where the bone material was slightly • The stone graves of Maidla I (fifth to sixth centu- burnt in comparison with that of the eastern tarands ries Ad) and II (tenth to 13th centuries Ad) were (Laul 2001, pp.42, 196). The analysed part of the cre- excavated by M. Mandel between 1983 and 1985, mains consists of 59 bone units. and between 1987 and 1990 (Mandel 2003, p.39). Suure Rõsna, Rõsna-Saare I and II sand-barrow cem- The first stone grave of Maidla was mainly used during eteries from the second half of the first millennium AD the fifth and sixth centuries AD, while some later buri- were excavated by Mare Aun during the 1980s (Aun als from the 12th and 13th centuries were also found 1992). The barrow cemeteries consist of rounded and (Mandel 2003, p.2). The bone material consists of long barrows of piled sand, with various burial cus-

122 BALTICA 19 BALTICA ARCHAEOLOGIA

SOCIETIES Of THE PAST: APPROACHES TO LAndSCAPE ig. 1. The Estonian cremation graves studied. f ig. 1.

123 toms; the cremated bones were commonly buried as fragment method. The most common well-preserved quite compact assemblages (sometimes in containers) bone fragment is a petrous part of the temporal bone. under and in the barrows (Aun 1992; Aun 2005). The The minimum number of buried people for the ana- human bone material collected from the barrows is all lysed cremations in Estonia was in most cases estimat- cremated (Allmäe, Maldre 2005; Allmäe et al. 2007). ed on the basis of this bone fragment. The total number of analysed bone units is 420. Percentage of determined bone finds in material. The Kirikumägi flat ground cemetery from the late tenth to percentage of bone units where the determination of the 12th century AD was excavated by Heiki Valk in at least one human bone fragment was calculated. The 2003 and 2004, and 2007. The scattered cremation bur- hypothesis is that this characteristic describes the dis- Observations on Estonian Iron Observations on Estonian Iron Age Cremations ials were detected at an underground level. A total of persal of bones: a high number of units with non-de- 69 bone assemblages were collected. The bones were termined bone fragments should define a burial place mainly cremated, with the exception of some unburnt with scattered cremation burials.

RAILI ALLMÄE teeth and temporal bones from a three or four-year-old child (Valk, Allmäe 2010). Colouration and fracture pattern on bones in different burial places Methodology Bone colour was not recorded according to accepted Bone assemblage or unit or find colour standards, as the material was analysed over quite a long period of time (1997 to 2011). during Usually, the cremated bone fragments are collected that time, the methodology for recording bone ma- from the burial site and packed as assemblages or terial changed to some extent. However, the general units. All these assemblages/units are numbered. Some notes and impressions of bone material are only briefly archaeologists call these units bone finds; some have mentioned, as some pattern was observable. The de- argued that this packed unit of bone fragments is not scription of changes in bone colour during cremation an archaeological find, but an assemblage of bones. It is discussed in the light of works by Shipman et al. does not matter what we call these packed bone units, (1984), McCutcheon (1992), Holck (2008), Walker et it is important to define what the unit, find or assem- al. (2008) and Wahl (2008). blage we study is. first, we have to realise that the bone ‘Charred’ bones occur in cremations when the oxy- assemblage/unit/find is not automatically one burial. gen flow and heat were insufficient for proper burning This is especially true of burial places where cremated (Holck 2008; Walker et al. 2008; Wahl 2008). Bone bones are scattered, and somewhat true for other types fragments indicating this appear as blackish or dark of grave where the burial technique is not recognised brown bone fragments among cremains: according on site. It all depends on the excavation methodology: to McCutcheon (1992), the specimens were heated to how the material was excavated, collected, packed and 340ºC; according to Holck (2008) grade 1 combustion described on site; or, to put it briefly, how well archae- (200–400ºC), to Wahl (2008) grade 2 (300–400ºC), ologists understand the bone material on site. On the and according to Shipman et al. (1984) stage II/III other hand, one packed and numbered collection of (285–525ºC). These bones often appear within the bone fragments is a good unit by which to study the group of bones which are burnt at higher temperatures peculiarity of the material, the fragmentations and the (Holck 2008, p.94). The appearance of black/dark dispersal of cremated bones in different graves. brown bones was recorded in the studied bone materials. The percentage of bone units with brown/blackish Minimum number of buried people bone fragments is presented as a percentage of the total and percentage of bone finds number of bone units within the studied material. determined The grey colour of bones appears at temperatures of 550ºC to 650ºC, according to Wahl (2008) and Walker Minimum number of buried people (MNI). It is known et al. (2008). Shipman et al. (1984) described light that some parts of the skeleton are more resistant to grey with secondary colours of brown and light brown- heat-induced changes, and preserve better among the ish grey from 525ºC to 645ºC. McCutcheon (1992) cremated bones (Holck 2008). Thus, the minimum observed light brownish grey at temperatures up to number of buried people among examined cremains is 600ºC. Holck (2008) determined dark and light grey very often determined according to the recurrent bone colours at temperatures of 400ºC to 800ºC. The grey

124 colour is attributed to bones which are not completely or afterwards. The maximum length of the long-bone burnt. fragment was measured for every packed bone unit for ten burial sites under examination. In addition, the At high temperatures of 800ºC or more, the bones be- biggest fragment of cranial vault was measured for ten come calcinated and the white colour of cremains is burial sites. Two measurements were taken for cranial prevalent, according to Holck (2008). McCutcheon vault fragments: the maximum length and the transver- (1992) and Wahl (2008) argue that the white colour sal length/width.The metric characteristic of the cra- appears at the somewhat lower temperature of 650ºC 19 BALTICA nial vault size is expressed as the cross multiplication to 700ºC. According to Shipman et al. (1984), the of these two measurements. A sliding calliper was used predominant colour of the bones is neutral white with for measuring bone fragments. blue-grey or light grey at temperatures of up to 940ºC. When bones are heated in excess of 940ºC, the specimens are neutral white with some medium grey and Radiocarbon dates of cremated human

reddish-yellow (Shipman et al. 1984). bones ARCHAEOLOGIA ‘Sooty’ bones occur in burials where the bones are The date of cremated bone fragments from Maidla I buried with pyre remains. This means that the bone (HELA-2403) and II (HELA-1958) stone graves, Rõs- fragments are not carefully picked out from crema- na-Saare I (HELA-1959; 1960) and II ( HELA-1961) tions and buried in the grave, or that the cremation site barrow cemeteries, Suure Rõsna (HELA-1962) barrow and the grave are the same, or that we have found the cemetery, and Põlgaste tarand grave (HELA- 2404; actual cremation site itself. Sooty bones are found in I 2405) was conducted in 2009 and 2010 at the dating all such cases. When clean bones occur in a grave, we Laboratory of the finnish Museum of natural History, SOCIETIES may conclude that the cremation pyre is located away Of THE PAST: University of Helsinki. The 14C concentration was from the grave (Sigvallius 1994). The occurrence of APPROACHES measured using the AMS method. The results are cali- TO LAndSCAPE sooty bones was recorded for the graves under exami- brated according to Intcal09 curve (Reimer et al. 2009) nation here. and Oxcal 4.1 software (Bronk-Ramsey 2009). fracture pattern of cremated bones Statistical analyses of descriptive The alteration of bones due to heat has been studied characteristics of different burial for decades. Changes in the surface structure of bones places are macroscopic and microscopic, and are different if The idea is to compare different graves on the basis dry bones or fleshed bones/corpses are cremated. The of correlations between characteristics observed. The thermal fractures and alterations characteristic of the average size of long-bone fragments, the proportion of cremation of corpses are warping, transverse and lon- determined bone finds among the material and the radi- gitudinal fractures, splintering and delaminating (often ocarbon dates of the graves should indicate differences seen on the cranial bone), curved transverse fractures, in burial practices. It should be possible to observe the patina, and so on (Buikstra, Swegle 1989; Symes et changes in burial practice over time, and also cultural al. 2008; Musgrave et al. 2010). Characteristic heat- and geographical differences in this practice. induced alterations of bones for burnt corpses were recorded for every grave. However, an inconsistent ob- R 2.15.2 was used for statistical analyses. servation of various patterns has to be admitted, as the pattern was not recorded for every bone unit. Results and discussion

Bone fragments and The analysed bone material is mainly from Middle their measurements Iron Age and Late Iron Age cemeteries (Aun 1992; Mandel 2003; Valk, Allmäe 2010). The only exception The most reasonable way to estimate the fragmentation is Põlgaste tarand grave, which is dated to the Early of cremains is to measure the biggest bone fragment(s) Iron Age (Laul 2001, p.40ff). in one packed unit. The bone fragment size was measured for all analysed materials. The bone fragment size is post-excavational, but it should be noted that the material was not sieved or washed during excavation

125 Radiocarbon dates minimum number of individuals differs from what is archaeologically predicted, and also from the osteolog- Radiocarbon dating (AMS method) of cremated bones ically determined plausible number of buried deceased. from six of the graves studied was conducted in or- The plausible number of burials is not discussed in the der to specify the time period when the grave was in present paper, although it has been described earlier use (Tables 1; 2). The main purpose of this analysis for west Estonian burial sites with scattered cremated was to find out if there is any correlation between the bones (Allmäe 2003; 2006). characteristics of the bone material and the time period The proportion of bone finds where at least one hu- when the grave was in use. The relative dates of the man bone fragment was recognised is also quite a

Observations on Estonian Iron Observations on Estonian Iron Age Cremations stone graves at Maidla are (Mandel 2003): the Maidla good characteristic to follow. The proportion of bone I grave, fifth to sixth century AD, and Maidla II, tenth finds determined is quite small for west Estonian Late

to 13th century Ad. The 14C dating of cremated bone Iron Age stone graves with scattered cremations (Ta- material confirms the gap, but indicates the use of the

RAILI ALLMÄE ble 2). The new form of cremation grave, Iron Age flat Maidla I grave somewhat earlier (Tables 1; 2). The ground burials, was detected in Kirikumägi, Siksälä, southeast Estonian sand barrow graves studied are dat- in southeast Estonia (Valk, Allmäe 2009; 2010). Here, ed to the sixth to the eighth centuries Ad (Aun 1992). as in west Estonian stone graves, the cremains were The radiocarbon dates indicate the use of sand barrow scattered and the number of bone units determined was cemeteries somewhat earlier (Tables 1; 2), and support relatively small. Thus, we may argue that the Late Iron the idea proposed by Mare Aun (1992, p.113, pp.151- Age cremation is distinguished from earlier sand bar- 152) that the barrows were piled up over a short time row cemeteries on the basis of this descriptive feature when an important member of the community died. (Table 2). The cremains in sand barrows of Long Bar- The tarand grave at Põlgaste was dated to the third to row Culture in southeast Estonia are buried under bar- the fifth century AD (Laul 2001, p.42ff). The AMS rows or in barrows, and the buried bone assemblages dates of cremated bones from the earlier part of the are mostly quite large and compact here (Aun 1992; grave here also refer to the somewhat earlier use of the 2005; Aun et al. 2008). Burial practices are quite eas- grave (Tables 1; 2). The AMS dates of the cremains ily distinguishable through the percentage of bone studied indicate somewhat earlier grave use in com- units determined. The same trend is observable for parison with former relative dates (Table 1). the Põlgaste tarand grave, the proportion of bone finds determined is highest here, and at the same time the minimum number of individuals is highest in relation number of bone finds, percentage of to the bone finds studied (Table 2). determined bone finds, and minimum number of buried individuals fracture pattern and colour of bones The number of cremated bone finds from west Estonia studied is 1,586, and from southeast Estonia 548. The Estonian cremations indicated a fracture pattern on the discrepancy is probably due to the burial practice that bones characteristic of the cremation of corpses (fleshed I will discuss below. Of course, a lot depends on how bone). The bone material from Põlgaste tarand grave we interpret the deposited cremains. If we conclude is somewhat different, where the absence of warping that every packed unit is one burial, then we calculate and longitudinal splitting, and only a few bones with an exceptional number of burials for graves with scat- curved transverse cracking, were observed. The same tered human cremations. On the other hand, burials in peculiarities have been registered at the Tandemägi IV sand barrows or in earlier tarand graves are easier to grave (Early Iron Age), where a secondary cremation understand, as the buried bone assemblages are com- burial after the flesh had decayed is supposed (Kal- pact here, and perceived and defined as burials already man 2000b; Lang 2007, p.180). Experiments with dry on site. bones have indicated that bones that are burned when dry show superficial checking and cracking, a lack of According to the method of recurrent bone fragments, longitudinal splitting, and no warping (Buikstra, Swe- the minimum number of buried individuals (MnI) in gle 1989, p.248). the west Estonian stone graves studied is 62, and in the graves of southeast Estonia 99. More detailed data It has been argued that there is no direct correlation is presented in Table 2. It is worth mentioning that the between the cremation temperature and the colour

126 Table 1. Radiocarbon dates (AMS method) of cremated bones

Grave Number Location Speci- Analysed Lab number Calibrated date of bone in grave men bone frag- radiocarbon find ment determination (95.4%) (68.2%) Maidla I MI-240 Square Adult? Long bone HELA 2403 258 (11,9%) 297Ad 341 Ad - 413 Ad 104f 1675 +-30 BP 321 (83,5%) 427 Ad BALTICA 19 BALTICA

Maidla II MII-180 Square Adult Cranial Hela-1958 980 Ad - 1060 Ad 980 Ad - 1060 Ad 96/96 n-o vault 1000 ± 30 BP

Põlgaste Po-2 Square Adult Cranial HELA 2404 40 BC (91,5%) 88 Ad 5 Ad - 74 Ad 8/b-c vault 1959 +-30 BP 103 Ad (4.4%)122 Ad

Põlgaste Po-3 Square Adult Cranial HELA 2405 45 BC - 80 Ad 20 BC (5,4%) 12 BC ARCHAEOLOGIA 10/f vault 1977+-30 BP 1 BC (62,8%) 65 Ad

Rõsna- - Long Adult Axis HELA-1959 410 Ad - 570 Ad 430 Ad - 540 Ad Saare I barrow 7 1565 ± 35 BP

Rõsna- RSI-157 Rounded Child Long bone HELA -1960 390 Ad - 550 Ad 420 Ad (29.3%) 470 Ad Saare I barrow 9 1595 ± 35 BP 480 Ad (38.9%) 540 Ad I

Rõsna- RSII-86 Long Adult Cranial HELA-1961 340 Ad - 540 Ad 390 Ad (36.3%) 460 Ad SOCIETIES Of THE PAST: Saare II barrow 7 vault 1620 ± 35 BP 480 Ad (31.9%) 540 Ad APPROACHES TO LAndSCAPE Suure- SR-40 Rounded Adult Cranial HELA-1962 420 Ad - 600 Ad 430 Ad (32.2%) 490 Ad Rõsna barrow 6 vault 1535 ± 35 BP 510 Ad (1.5%) 520 Ad 530 Ad (34.5%) 580 Ad

of the bones, and that the colour should be observed jects only some general notes and observations are pre- together with alterations in the bone structure (Ship- sented. man et al. 1984; Holck 2008; Walker et al. 2008; Wahl Dark brown and black bone fragments were occasion- 2008). Sometimes, a wide range of colour alterations is ally found in nearly all the Estonian graves studied found within a single cremation. This is especially true (Table 2). The appearance of incompletely burnt dark when whole corpses/fleshed remains have been burnt brown and black bones within the group of bones, (Symes et al. 2008, p.35ff; McKinley 2008, p.168ff). which are burned at higher temperatures, has been re- Earlier experiments and research that show that the corded earlier (Holck 2008, p.94). Bone material from colour of the bones alone is not a trustworthy indica- other parts of Estonia also contains heat-altered bone tor of the burning temperature are convincing. In ad- fragments, which are characteristic of the grade 1 burn- dition, it should also be noted that the colour of burnt ing described by Holck (2008); for example, cremated bones may change if they are buried (Shipman et al. cranial fragments in Viimsi tarand grave, from 350 to 1984; Wahl 2008). Shipman et al. (1984) also mentions 500 Ad, in northern Estonia (Kalling 1993; Lang 1993, errors in the determination of colour, because of indi- p.55; Lang 2007, p.133; Kivirüüt 2011). Here we can vidual differences in the research used to perceive fine find glassy black and dark brown fragments of cranial colour distinctions. vault very similar at first sight to pieces of ceramics. Additionally, in the Tõnija tarand grave excavated by for the Estonian bone materials investigated, colour M. Mägi between 1995 and 2001 (Mägi 2001), the standards were not used to determine the colour of the author of the present paper observed dark-brown and cremains, the observations on fracture pattern were not consistent, and microscopic investigations on bone structure were not conducted. Therefore, for these sub-

127 Table 2. The dates and descriptive characteristics of analysed cremation graves

West Estonian grave Southeastern Estonian graves

name of Uugla Rõsna- Rõsna- Suure Kiriku- Kirbla Ehmja Maidla I Maidla II Põlgaste grave III Saare I Saare II Rõsna mägi number of studied bone 70 248 268 674 326 59 159 97 164 69 Observations on Estonian Iron Observations on Estonian Iron Age Cremations units

Stone- Stone- Stone- Stone- Stone- Tarand- Sand Sand flat- Grave type Sand barrow grave grave grave grave grave grave barrow barrow ground RAILI ALLMÄE Relative dates 11-12 11-12 5-6 10-13 11-13 3-5 6-8 7-8 6-8 - (cc.) (5-7) (12-13)

Radiocarbon 258-427 980-1060 40BC-122Ad 410-570 Ad 340-540 420-600 990-1160 - - - dates (95,4%) Ad Ad 45BC-80Ad 390-550 Ad Ad Ad Ad* probability % of determined bone 13.0 17.2 41.9 29.7 18.1 81.7 70.6 66.6 55.1 25.6 units

% of bone units with ‘charred’ 2.6 1.8 9.2 9.7 5.8 3.4 11.4 9.4 4.8 2.6 bones

Minimum number of 1 5 17 32 7 18 29 22 26 4 burials

* radiocarbon dated charcoal amongst cremated bones. The other date is somewhat earlier (68.3% probability 723-893 Ad), but no bones were found with charcoal here (Valk, Allmäe 2010)

brownish-black fragments of human cranium among graves, with the largest grave area and the largest the unburned human bones. The different handling of number of cremation burials (fig. 1; Table 2). The head and body is observed in tarand graves, and is as- cremains in west Estonian stone graves were usually sumed to be a ritual practice (Lang 2007, p.180). pale, often with white and grey dominating, sometimes with brownish and blue hues, so the temperature of the cremation pyre should have been at least moder- West Estonian stone graves ate. for example, in the Uugla III stone grave, 87.1% of the bone finds are described as white, with a blue Kirbla, Ehmja and Uugla III are quite small cemeter- hue also being very frequent. for the Maidla II grave, ies, containing seven or fewer burials. The Maidla I Mandel (2003, pp.42ff, p.59) has described the bone grave is somewhat larger: the minimum number of fragments as very burnt, and observed the occasional buried individuals was 17. The Maidla II stone grave melting of bronze artefacts in the grave. The Maidla I is most representative of the studied west Estonian stone grave contained slightly and heavily burnt bones

128 (Mandel 2003, pp.30-40), and the bone fragments here sand barrows (Suure Rõsna, Rõsna-Saare I and II). The were also slightly bigger in comparison with the other prevailing colour of the bone units in sand barrows west Estonian graves studied (Allmäe 2003, p.250ff). was white with a yellow hue (39.5% of bone units), The bone fragments were mainly a pale colour with with the next most common colour described as pale a brownish hue. McCutcheon (1992) observed a light with a greyish beige or greyish yellow hue (37.1%), brownish grey in a specimen heated to 600ºC. Wahl and a brownish or dark grey colouration recorded for (2008) has observed white with brownish, greyish and 10.2% of bone units. Dark brown and black colours ochre colours at a temperature of 800ºC and above, and (incompletely burned bones) were recorded for 8.6% 19 BALTICA argues that these secondary colours correspond to local of bone units, while bone units buried with charcoal soil conditions. and soot (‘sooty bones’) formed 11% of buried bone assemblages. The proportion of bone finds with a dark brown and/ or black colour among west Estonian graves was more Bone units consisting of white fragments with a yellow

frequent in the Maidla stone graves (Table 2). hue were frequent, suggesting that these bones were ARCHAEOLOGIA ‘clean’, not buried with charcoal or soot, but carefully Sooty bones were quite common in west Estonian picked from the pyre remains before burial. For exam- stone graves too, as the cremation often took place in ple, ‘clean’ bones form 30.2% of bones from Swedish the grave area (Mandel 2003, p.153ff; Allmäe 2003). Early Iron Age cremations, and only 0.8% of Late Iron Age cremations (Sigvallius 1984). On the other hand, Southeast cremation burials many bone assemblages (11%) were pyre remains, I buried in sand barrows with a remarkable amount of The studied southeastern graves with cremations are charcoal and soot. The occurrence of ‘clean and white’ SOCIETIES Of THE PAST: of varied construction: Põlgaste is a tarand grave, Kiri- and ‘greyish brown and sooty’ bones in sand barrows APPROACHES kumägi is a flat ground cemetery, and others are sand demonstrates different contemporaneous burial prac- TO LAndSCAPE barrow cemeteries (fig. 1; Table 2). tices of cremains in southeast Estonia. The prevailing colour of bones within the Põlgaste The surface patterns on bones indicate typical altera- tarand grave material is pale with a light brown and oc- tion, warping, splitting and cracking caused by heat. casional orange-red hue. Shipman and her colleagues The colour of the bones gives us a hint of the heat of (1984) have described reddish-brown and reddish-yel- the cremation pyre, and also of the burial environment. low colours on burnt bones at temperatures of 285ºC White bones with a yellow hue are evidence of quite a to 645ºC. The archaeologist Silvia Laul (2001, p.196), high pyre temperature, at least 800ºC. A yellow hue on who excavated Põlgaste tarand grave, has argued that the surface of the bones and in bone spongiosa is quite the bones were slightly burnt. common for cremations above 800ºC (Wahl 2008). The cremated bone material in the Põlgaste grave ex- The greyish beige or greyish-yellow colour is much hibited the occurrence of calcinated fragments and sur- more difficult to interpret. The suggestion is that the prisingly uniform colouration. The latter is attributed to bones are moderately burnt, probably at temperatures the cremation of dry bones, here calcinated bones are between 600ºC and 800ºC, with a partial greyish tone light brown or tan (Buikstra, Swegle 1989, p.249ff). on fragments caused by incomplete oxidation. A grey and light-grey colour of bones is observed for a wide Dark brown and black fragments were very rarely range of pyre temperatures, mainly beginning around observed among the material, while sooty bone finds 550ºC to 650ºC; dark greyish-brown bone units are were absent (Table 2). probably cremated at temperatures lower than 600ºC In Kirikumägi flat ground cemetery, mainly light- (McCutcheon 1992; Shipman et al. 1984; Holck 2008; coloured bone assemblages were found, with shades Walker et al. 2008; Wahl 2008). varying from light yellow-greyish to light-brownish; Observations on the colouration of cremated bones occasionally dark grey shades were also observed, indicate varied cremation temperatures in nearly all The cremation temperature should have been at least the graves studied. The exceptions are Põlgaste tarand 600ºC, most likely above 800ºC (Valk, Allmäe 2010). grave and Uugla III stone grave, where the colouration More detailed observations were conducted to describe of the bone material was quite uniform. the colouration of the cremains in southeast Estonian

129 Bone fragmentation and burial some hypotheses based on bone fragmentation and practice burial customs in the graves studied are discussed below. during the analysis of Estonian cremations, the maxi- The scattering of the cremains of several people over mum length of long-bone fragment and the maximum the grave causes the dispersal of bone fragments in a size of cranial vault fragment were measured in order way that makes the determination of the number of in- to analyse the overall fragmentation of the grave mate- dividuals in the grave difficult, and causes high num- rial. Both measurements show a decline with time (Ta- bers of small non-determined bone units. On the other bles 3; 4; fig. 2). Põlgaste tarand grave is exceptional hand, burials with compact nested bone assemblages,

Observations on Estonian Iron Observations on Estonian Iron Age Cremations among the materials studied, as here most likely dry where fragments are carefully picked up from the pyre bones were cremated. site, of course indicate less fragmentation and a higher The overall fragmentation of the buried cremated proportion of identified individuals in graves (Tables 2; bones is caused by the heat of the funeral pyre and the

RAILI ALLMÄE 3; 4; fig. 2). The bone fragment size of modern crema- handling of the cremains afterwards: cooling, raking, tions is comparable to the ancient cremations buried in collecting, transporting, crushing and burying may well-closed containers (McKinley 1994). cause additional fragmentation (Lange et al. 1987; The bone materials from Late Iron Age graves are McKinley 1994; 2008; Sigvallius 1994; Holck 2008). more fragmented and indicate a smaller proportion of The further fragmentation of cremains can be the result determined bone fragments than southeast Estonian of excavating the bone material, then sieving, cleaning graves (Tables 3; 4; Fig. 2). We may assume that this is and washing the bone material during post-excavation caused by the mortuary practice of the Late Iron Age, processes, as well as when depositing and analysing scattering cremains in the grave area, although we can- the bone material. It has been argued that the excava- not rule out the possibility that the bones were crushed tion and the post-excavation processes can cause the before burial, or that some bone parts were taken from largest amount of damage to cremains (Lange et al. the pyre site and then buried somewhere else. How- 1987; McKinley 1994; formisto 1996). ever, can we always distinguish the pyre site from the We may hypothesise that, besides the processes dur- burial site or cult place? ing cremation, and during excavation and post-exca- One burial place in Estonia where very few cremated vation activities, the fragmentation stage of cremains fragments of human bone were detected is the Viking describes some other factors too. for example: 1) the Age stone circle graves in Piila, on the island of Saare- burial custom itself, scattered cremations or compact maa. Here, among highly fragmented calcinated ani- nested burial of cremains, sometimes in some kind of mal bones, only a few fragments of human bones were container; 2) the variability of actions with cremains determined (Mägi et al. 1998). Graves with few human before burial, sorting, collecting and even crushing bones, or a total absence of them, have been described bones before funeral rituals. earlier, often along with the reasoning of distribution Excavation and post-excavation activities are easily of cremains between several (burial or ritual) places detectable, as the activities after the funeral pyre and (McKinley 1989; Kaliff 1992, p.121ff; Sigvallius before burial are in a ‘grey area’ for ancient cremations. 1994, p.27ff; Kaliff, Oestigaard 2004; Williams, 2008; We do not know how our ancestors thought and acted Parker Pearson 1999, p.55); thus the concept of ‘grave’ during the death and funeral of a family or community can be symbolic. member. The reason is that we do not understand their Arcini (2005) has argued that pyre sites contain a rela- beliefs, and we will never have a clear vision of their tively large number of animal bones, or even whole burial customs. Therefore, partial burials of cremated skeletons, while at the same time very few human bones, or burying bones in several places, or crushing bones are detected. furthermore, another characteristic bones before burial, are highly speculative subjects to trait of osteological material from pyre sites is the lack handle, because these activities are not easily detect- of the petrous parts of the temporal bone. able by archaeological and osteological methods. This is definitely not the case for southeastern graves, despite the argument that activities after the funeral because compact bone assemblages are buried here, pyre and before the burial are quite difficult to detect, the size of the bone fragments is large, and the propor-

130 BALTICA 19 BALTICA ARCHAEOLOGIA

SOCIETIES Of THE PAST: APPROACHES TO LAndSCAPE ig. 2. The correlation between radiocarbon date, bone fragmentation and the proportion of determined units in f ig. 2. grave.

131 Table 3. Long-bone fragment maximum length (average per grave) and radiocarbon dates of graves

Sample Kirbla Ehmja Maidla I Maidla II Uugla III Põlgaste Rõsna- Rõsna- Suure Kiriku- Saare I Saare II Rõsna mägi dates 11-12 5-7/11-12 4-5 10-11 11-13 Ad* 1 BC-1 5-6 4-6 5-6 11-12 cc. Ad* Ad* Ad Ad Ad Ad Ad Ad Ad n 51 151 205 541 316 54 106 61 140 29 Ave 3.2 2.2 4.3 3.2 2.7 5.7 4.7 4.9 4.3 3.0 Mode 3.0 2.0 3.5 2.2 2.7 4.5 4.4 3.8 5.1 1.3

Observations on Estonian Iron Observations on Estonian Iron Age Cremations Median 3.0 2.0 3.8 3.0 2.7 5.4 4.4 4.6 4.0 2.4 Max 5.0 8.6 10.1 17.2 5.2 12.6 9.5 9.8 8.8 6.6

Min 1.7 0.9 1.7 1.1 0.8 1.8 1.8 1.7 1.2 1.1 STd 0.8 1.0 1.5 1.3 0.8 2.6 1.8 1.8 1.6 1.6 RAILI ALLMÄE * Relative dates

Table 4. Cranial vault fragment maximum size (average per grave) and radiocarbon dates of graves

Grave Kirbla Ehmja Maidla I Maidla II Uugla III Põlgaste Rõsna- Rõsna- Suure Kiriku- Saare I Saare II Rösna mägi dates 11-12 5-7/11-12 4-5 10-11 11-13 1 BC-1 5-6 4-6 5-6 11-12 cc. Ad* Ad* Ad Ad Ad* Ad Ad Ad Ad Ad

n 27 75 139 248 192 47 65 40 84 15 Ave 3.4 2.7 7.4 5.0 3.2 12.9 8.6 8.9 7.3 4.6 Mode 8.6 1.0 7.1 2.2 2.4 5.5 8.4 - 1.9 4.6 Median 2.9 2.0 5.9 4.1 2.9 12.0 7.8 6.9 6.1 4.6 Max 8.6 15.8 28.1 32.7 11.3 31.3 23.4 42.3 23.6 11.0 Min 0.9 0.5 1.1 0.9 0.7 3.6 1.7 2.1 1.5 1.3 STd 2.3 2.7 5.0 3.8 1.5 6.4 5.2 7.6 4.8 2.7

* Relative dates

tion of determined bone units and minimum number in addition to the pyre sites, an enormous amount of of burials are high (Tables 2; 3; 4; fig. 2). It should human cremains were found over a large area. also be noted that pyre sites have not been found near In Laitila in finland, the Merovingian (600–800 Ad) southeastern cremation cemeteries. The cremations cremation cemetery A, at Vainionmäki, was exca- themselves took place somewhere else, and the cre- vated between 1986 and 1994. The average size of mains were collected and carried to the burial place. the longest bone fragments measured in grave A was We may speculate that the Late Iron Age graves at one to three centimetres, with a maximum length of Kirbla, Uugla I and III are pyre sites, because burnt about eight centimetres. The bone colour is described areas with greasy black soil, often with small pieces as grey, although the bones were covered with soot, of artefacts and heavily cremated bones, were found which sometimes caused difficulties in determining the (Mandel 2003, pp.77, 88-93; Mandel, Allmäe 2009). actual colour (formisto 1996, pp.81-87). despite the The petrous parts of the temporal bones in the Uugla time gap, the Vainionmäki A cremation grave, with ir- graves were found mainly in pieces, not even a piece regular stone settings, scattered cremains, often-found was found in the Kirbla grave (Allmäe 2003, pp.246- sooty bones, and also with a comparable fragmentation 247; Mandel, Allmäe 2009). However, this explanation stage of bones, is very similar to west Estonian graves, does not fit the Maidla I and II stone graves, because, especially to the Maidla I stone grave of the fourth to

132 fifth centuries, as we will see below. Formisto (1996, support the idea that the cremated remnants of children pp.81-87) has suggested for Vainionmäki that the cre- disappear in the grave with time (Holck 2008, p.119). mation itself was conducted in the grave area. for sev- The idea of crushing cremains is supported by some eral west Estonian cremation graves, Mandel (2003) researchers, while others disagree. In the middle of has suggested the same, and has described heavily the last century, the deliberate crushing of bones after burnt areas with a greasy charcoal layer in the grave cremation was proposed by Gejvall (1959). Sigvallius area. This is in accordance with criteria established by

(1994, p.27ff) noticed that cremations from 0 to 500 19 BALTICA other researchers earlier (Iregren 1972, p.73; McKin- Ad (Pre-Roman Iron Age in Sweden) are much better ley 1989) to determine the areas where corpses were preserved in comparison with Viking Age graves, and burnt in the grave area. has supposed that crushing cremated bones is highly Thus, it is probable that in west Estonian cremation likely. McKinley (1994, p.339) argues that the frag- graves, the bodies were burnt in the grave area, and mentation of bones is caused mainly by excavation

the cremains were scattered over the area or partially and post-excavation activities. formisto (1996, p.86ff) ARCHAEOLOGIA buried somewhere else. However, this does not explain agrees with McKinley on this point, and also says that the variation in the number of child cremations, as we the fragments found in different cremation graves are shall see below. almost always the same. It should be noted that white or pale very burnt bones The author of the present paper tends to support the from southeastern sand barrows are very fragile: crush- idea of Scandinavian research at the moment, for two ing the spongy parts of the bones is easy. In west Est- main reasons. first, the investigated Estonian cremated I onian Late Iron Age graves, the spongy parts of the bone material was not sieved or washed during archae- SOCIETIES bones are found less, as are the cremains of children. ological excavations (except at the Uugla III grave). Of THE PAST: APPROACHES The proportion of child cremations in the Maidla II Secondly, it must be emphasised that the Estonian bone TO LAndSCAPE grave from the Late Iron Age (4.7%) is comparable material was also not washed or sieved before osteo- to those in Scandinavian cremations (Sigvallius 1994; logical analysis, the material is post-excavational, but Holck 2008, pp.63, 119). The exception is the Maidla not handled due to the standardised procedures for cre- I grave from the fourth and fifth centuries, where the mated bones. Among deposits, an amount of soil, char- cremains of six children (35.3% of the total number coal and pyre debris is often found. Thus, our material of cremations) were found (Allmäe 2003; 2006). The has a somewhat different character. Therefore, we may Finnish Merovingian period grave A at Vainionmäki speculate that the crushing of cremated bones has been contained 23 cremations, of which seven (30.4%) were practised, or the bodies of children were handled dif- of children (formisto 1996, p.81ff). In these two graves ferently in Late Iron Age Estonia. Of course, further with scattered cremains, the proportion of child cre- studies to support this working hypothesis should be mations is similar. The cremated remains of children carried out. are very fragile (Holck 1995; Holck 2008, p.119ff), crushing them before burial makes them invisible in Conclusions graves. On the other hand, the cremains or bodies of infants and children could have been buried elsewhere. The proportion of determined bone units is different for example, in western Estonia, Late Iron Age infant between the 900 to 1250 Ad and the 40 BC to 600 Ad inhumations have been found in Bronze Age stone-cist graves, the latter indicating higher proportions. The graves (Allmäe 2010). graves from 40 BC to 600 Ad also indicated less bone The sand barrow cemeteries in southwest Estonia in- fragmentation. Both measurements (cranial and long- dicated a different mortuary practice concerning chil- bone fragment) decline with time; the earlier graves dren. Cremations of infants and children were quite with burials of compact bone assemblages show a commonly found here amongst adult burials; for exam- lower stage of fragmentation than the later ones. This ple, in the Rõsna-Saare I grave, the proportion of child also confirms the fact that the custom of scattered cre- burials was 37.8% (Allmäe, Maldre 2005; Allmäe et mation burials is distinguishable on the basis of higher al. 2007; Allmäe et al. 2009). The results of the analy- bone fragmentation. sis of southeast Estonian barrow cemeteries definitely nearly all bone materials indicated fracture patterns rules out the custom of crushing the bones before bur- and surface patterns characteristic of burning corpses, ial, or of burying children and infants other than in the the only exception being the tarand grave at Põlgaste family/community cemetery. In addition, this does not from the beginning of the first millennium. Here, a pos-

133 sible secondary burial is suggested, the cremation of References bones. The colouration of cremated bone material varied within the graves. Temperatures of between 600ºC and 800ºC are suggested for funeral pyres if the bones Manuscripts are a pale colour. nearly all the graves indicated in- KIVIRÜÜT, A., 2011. Põletatud luude uurimine: me- completely burnt bone fragments of a dark brown and toodika ning praktika Viimsi I tarandkalme leiukom- black colour; the proportion was higher for the Maidla pleksi näitel. Bakalaureusetöö. Tartu: Tartu Ülikool. stone graves in western Estonia, and for the southeast Available from: http://www.arheo.ut.ee/docs/BA11_ Estonian sand barrow cemeteries. The Põlgaste tarand Kivir%C3%BC%C3%BCt.pdf (Acces-ed 6.01.2012)

Observations on Estonian Iron Observations on Estonian Iron Age Cremations grave and Uugla III stone grave both indicated uniform colouration of bone material. Literature

The burial practices in western and southeastern Es- ALLMÄE, R., 2003. Läänemaa 5.–13. sajandi kalmete an-

RAILI ALLMÄE tonia were different. Pyre sites are detected in west tropoloogiline aines. In: M. MAndEL, ed. Läänemaa Estonian graves, thus cremation probably took place 5.–13. sajandi kalmed. Appendix 1. Tallinn: Eesti Aja- on the burial site, and most of the cremains were scat- loomuuseum. Töid Ajaloo Alalt 5, 243-262. ALLMÄE, R., MALdRE, L., 2005. Rõsna-Saare I tered nearby. The crushing of bones before burial is not kääbaskalmistu - esialgseid osteoloogilisi andmeid. Setu- observable for southeast barrow cemeteries, but it is maa kogumik. 3 : Uurimusi Setumaa loodusest, ajaloost not ruled out for Late Iron Age graves. The presence ja folkloristikast = Setumaa symposium. 3 : research into of child cremations is often higher for southeastern nature, history and folklore of Setumaa = Setumaaskii graves. The fourth to fifth-century Maidla I grave also sbornik. 3 : issledovaniaa po prirode, istorii i folkloristike Setumaa. Tallinn: Ajaloo Instituut, 121-137. indicated a high proportion of child cremations. for the ALLMÄE, R., 2006. Grave 2 of Maidla – the burial site of Late Iron Age Maidla II stone grave, the crushing of a single family. Estonian Journal of Archaeology 10 (1), bones before burial or different burial customs for chil- 3-23. dren and infants is suggested, as child cremations in ALLMÄE, R., AUn, M., MALdRE, L., 2007. Cremations the grave were extremely rare. The other Late Iron Age of the Culture of Long Barrows in northern Setumaa in The Second Half of the first Millennium. Preliminary re- graves were too small to draw this kind of conclusion. sults. Humanbiologia Budapestinensis, 30, 113-122. The change in mortuary practice that we can perceive ALLMÄE, R., AUn, M., MALdRE, L., 2009. Rõsna-Saarõ through different bone fragmentation stages, and in I kääbaskalmistu. In: M. AUn, ed. Setomaa 2. Vanem ajalugu muinasajast kuni 1920. aastani. Tartu: Eesti Rah- proportion to the determined bone units, is temporal on va Muuseum, 88-94. one hand, as in the later graves with scattered crema- ALLMÄE, R., 2010. Some remarks on Kaseküla stone-cist tions, the average, as well as the median size of bone grave, Läänemaa, Estonia. Fennoscandia Archaeologica, fragments, is smaller. On the other hand, the difference XXVII, 45-52. in mortuary practice, which is observable through the ARCInI, C., 2005. Pyre sites before our eyes. In: T. AR- TELIUS, f. SVAnBERG, eds. Dealing with the dead. presence of infant and child cremations in graves, is Archaeological perspective on prehistoric Scandinavian cultural as well as temporal. To be more precise, the burial ritual. Stockhohn, 63 72. cremains of infants and children are often found in the AUN, M., 1992. Arkheologicheckie pamiatniki vtoroi graves of Long Barrow Culture in southeastern Estonia polovinu 1-go tyciacheletiia n.e. v Iugo-Voctochnoi Esto- in the fourth to sixth centuries, and are well represent- nii. Tallinn: Olion. AUN, M., 2005. Pikk-kääbaste ehitusest. Setumaa kogumik. ed in the fourth to fifth-century stone grave of Maidla I. 4: Uurimusi Setumaa loodusest, ajaloost ja folkloristikast To support these hypotheses, the necessity to analyse = Setumaa symposium. 4 : research into nature, history and folklore of Setumaa = Сетумааский сборник. 4 : more cremated bone materials of various dates and исследования по природе, истории и фолклористике from different cultural areas is obvious. Сетумаа. Tallinn: Ajaloo Instituut, 97-120. AUN, M., ALLMÄE, R., MALDRE, L., 2008. Pikk-kääbaste tähendusest (Rõsna küla kääbaskalmistuste materjali põh- Acknowledgements jal). In: Ü. TAMLA. Setumaa Kogumik 4. Tallinna Ül- ikooli Ajaloo Instituut: Tallinn, 269 -290. This study was undertaken within the framework of the BROnK RAMSEY, C., 2009. Bayesian analysis of radiocar- target funded research project of the Estonian govern- bon dates. Radiocarbon, 51, 337–360. BUIKSTRA, J.E., SWEGLE, M., 1989. Bone Modification ment (Sf0130012s08). due to Burning: Experimental Evidence. In. R. BOnnIS- CHEn, M.H. SORG, eds. Bone modification. Orono: Peo-

134 pling of the Americas Publications, Center for the Study MAndEL, M., ALLMÄE, R., 2009. Ergebnisse der of the First Americans, Institute for Quaternary Studies. Archäologischen Ausgrabungen in Uugla. Arheoloogilised University of Maine, 247-258. välitööd Eestis = Archeological Fieldwork in Estonia fORMISTO, T., 1996. Osteological analyses. In: P. 2008, 114-123. PURHOnEn, ed. Vainiomäki – A Merovingian Period MCCUTCHEOn, P., 1992. Burned archaeological bone. In: Cemetery in Laitila, Finland. Helsinki: National Board of J.K. STEIn, ed. Deciphering a Shell Midden. San diego, Antiquities. Ca: Academic Press, 347-368. GEJVALL, N.-G., 1959. Vanligaste ben: Nagot om bestämn- MCKInLEY, J., 1989. Cremations: expectations, method-

ing av brända ben och deras vetenskapliga värde. Fynd. ologies and realities. In: C.A. Roberts, f. Lee, J. Bintliff, 19 BALTICA Göteborgs och Bobus läns tidskrift, 40-47. eds. Burial Archaeology: Current Research, Methods and HEAPOST, L., 2007. The cemetery of Siksälä osteological Developments. London: BAR British Series, 65-76. and paleodemographical analysis. Siksälä. In: S. LAUL, MCKInLEY, J., 1994. Bone fragment Size in British Cre- H. VALK, eds. Community at the Frontiers. Iron Age and mation Burials and its Implications for Pyre Technology Medieval. Tallinn-Tartu:Tartu University, 213-236. and Ritual. Journal of Archaeological Science, 21(3), 339- HOLCK, P., 1995. Why are small children so seldom found 342. in cremations? Cremation Studies in Archaeology. In: MCKINLEY, J., 2008. In the heat of the pyre: efficiency ARCHAEOLOGIA E. SMITS, E. IREGREn & A. G. dRUSISnI, eds. Pro- of oxidation in Romano – British cremations – did it re- ceedings of the Symposium Amsterdam 26.-27. October ally matter? In: C.W. SCHMIDT & S.A. SYMES, eds. The 1995. Amsterdam, 33-38. Analysis of Burned Human Remains. London: Academic HOLCK, P., 2008. Cremated Bones. A Medical-Anthroplog- Press, 162-184. ical Study Of an Archaeological Material on Cremations MUSGRAVE, J., PRAG, A.J.N.W., NEAVE, R., FOX, R.L., Material. 3rd edition. Oslo: University of Oslo. WHITE, H., 2010. The Occupants of Tomb II at Vergina. IREGEN, E. 1972. Vårby och Vårberg II : studie av kremerat Why Arrhidaios and Eurydice must be excluded. Interna- människo- och djurbensmaterial från järnåldern. Theses tional Journal of Medical Science; 7(6), 1-15. I and Papers in North-European Arcaheology. Stockholm: MÄGI, M., ALLMÄE, R., MALDRE, L., 1998. Viking Age SOCIETIES University of Stockholm. graveyard at Piila, Saaremaa. Arheoloogilised välitööd Ee- Of THE PAST: JAAnITS, L., LAUL, S., LÕUGAS, V., TÕnISSOn, E., stis 1997. Muinsuskaitseamet: Tallinn, 99-116. APPROACHES 1982. Eesti esiajalugu. Tallinn: Valgus. MÄGI, M., 2001. Probable cult site beside the Tõnija tarand- TO LAndSCAPE KALIFF, A., 1992. Brandgravskick och föreställningsvärld. grave on the Island of Saaremaa. Archaeological Field- En religionsarkeologisk diskussion. In: Occasional Papers work in Estonia = Arheoloogilised välitööd Eestis 2000. in Archaeology, 4. Uppsala: Societas Archaeologica Up- Tallinn: Muinsuskaitseamet, 48 55. saliensis. nICHOLSOn, R., 1993. A Morphological Investigation KALIff, A., OESTIGAARd, T., 2004. Cultivating Corpses. of Burnt Animal Bone and an Evaluation of its Utility in A Comparative Approach to disembodied Mortuary Re- Archaeology. Journal of Archaeological Science, 20 (4), mains. Current Swedish Archaeology, 12, 83- 04. 411-428. KALLING, K., 1993. Viimsi kalmete luuainese antropoloo- PARKER PEARSOn, M., 1999. The Archaeology of Death giline analüüs. In: V. LAnG, ed. Kaks tarandkalmet Viim- and Burial. Sutton: Phoenix Mill. sis, Jõelähtme kihelkonnas. Tallinn, 67-69. REIMER, P.J., BAILLIE, M.G.L., BARd, E.A., BAYL- KALMAN, J., 2000a. Stone grave II of Tõugu – skeletal re- ISS, E., BECK, J.W., BLACKWELL, P.G., BRONK port. Muinasaja teadus,7, 387-402. RAMSEY, C., BUCK, C.E, BURR, G.S., EDWARDS, KALMAN, J., 2000b. Tandemägi stone grave – osteological R.L., fRIEdRICH, M., GROOTES, P.M., GUILdER- report. Muinasaja teadus, 7, 423-434. SOn, T.P., HAJdAS, I., HEATOn, T.J., HOGG, A.G., KALMAN, J., 2000c. Uusküla II skeletal analysis. Muina- HUGHEn, K.A., KAISER, K.f., KROMER, B., MC- saja teadus, 7, 437-440. CORMAC, F.G., MANNING, S.W., REIMER, R.W., LAnG, V., 1993. Kaks tarandkalmet Viimsis Jõelähtme ki- RICHARdS, d.A., SOUTHOn, J.R., TALAMO, S., helkonnas. Tallinn. TURNEY, C.S.M., van der PLICHT, J., WEYHENMEY- LAnG, V., 2007. The Bronze Age and Early Iron Age in Esto- ER, C.E., 2009. IntCal09 and Marine09 Radiocarbon Age nia. Estonian Archaelogy 3. Tartu: Tartu University Press. Calibration Curves, 0–50,000 Years cal BP. Radiocarbon, LANGE, M., SCHUTKOWSKI, H., HUMMEL, S., HER- 51 (4), 1111-1150. RMAnn, B., 1987. A bibliography on Cremation. PACT SHIPMAn, P., fORSTER, d. & SCHOEnIGER, M., 1984. 19. Rixensart: Hackens. Burnt bones and teeth: an experimental study of colour, LAUL, S., 2001. Rauaja kultuuri kujunemine Eesti kaguo- morphology, crystal structure and shrinkage. Journal of sas. Tallinn: Ajaloo Instituut - Õpetatud Eesti Selts. Archaeological Sciences, 11, 307-325. MALDRE, L., 2003. Läänemaa kivikalmte arheozooloo- SIGVALLIUS, B., 1994. Funeral Pyres. Iron Age Crema- giline materjal. In: M. MAndEL, ed. Läänemaa 5.-13. sa- tions i North Spånga. Theses and Papers in Osteology I. jandi kalmed. Appendix 2. Tallinn: Eesti Ajaloomuuseum, Stockholm. Töid Ajaloo Alalt, 5, 243-263. SYMES, S.A., RAINWATER, C.W., CHAPMAN, E.N., MANDEL, M., 2003. Läänemaa 5.-13. sajandi kalmed. Tal- GIPSOn, d.R., PIPER, A.L., 2008. Patterned thermal de- linn: Eesti Ajaloomuuseum, Töid Ajaloo Alalt ,5, 5-197. struction of human remains in a forensic setting. In: C.W.

135 SCHMIdT, S.A. SYMES, eds. The Analysis of Burned 1 586 kremuoti kaulai, Pietrytinėje Estijoje – 548. Šie Human Remains. London: Academic Press, 15-54. kaulių kiekiai atitinka minimalų 62 ir 99 individų skai- VALK, H., ALLMÄE, R., 2009. Põletusmatused Siksälä čių. Procentiškai skiriasi 900–1250 AD ir 40 BC – 600 Kerigumäel. Setomaa, 2. Vanem ajalugu muinasajast kuni 1920. aastani. Tartu: Eesti Rahva Muuseum. AD laikotarpiais datuotų kapų kaulų proporcijos, t. y. VALK, H., ALLMÄE, R., 2010. Kirikumägi at Siksälä: evi- vėlesniuose kapuose kaulų rasta daugiau. 40 BC – 600 dence of a new grave form of South-Eastern Estonia. Esto- AD kapuose taip pat nustatyta mažesnė kaulų fragmen- nian Journal of Archaeology, 14, 40-55. tacija. Kaukolės ir ilgųjų kaulų fragmentų matavimai WAHL, J., 2008. Investigations on Pre-Roman and Roman rodo, kad kaulų mažėja laikui bėgant – ankstesnieji Cremation remains from Southwestern Germany: Re- sults, Potentialities and Limits. In: C.W. SCHMIDT, S.A. kapai su kompaktiškais kaulų rinkiniais rodo buvus že-

Observations on Estonian Iron Observations on Estonian Iron Age Cremations SYMES, eds. The Analysis of Burned Human Remains. mesnį jų fragmentacijos lygį, nei vėlyvesniuose kapuo- London: Academic Press, 145-162. se randami kaulai. Šis tyrimas patvirtina, kad paprotys WILLIAMS, H., 2008. Towards An Arcaheology of Crema- paskleisti degintinius kaulus yra atskiriamas, remiantis tion. In: C.W. SCHMIDT, S.A. SYMES, eds. The Analy- didesne degintinių kaulų fragmentacija. sis of Burned Human Remains. London: Academic Press, RAILI ALLMÄE 239-269. Beveik visų tirtų kaulų paviršius yra suskilinėjęs, o tai būdinga degintiems palaikams, Vienintelė išimtis Received: 30 november 2012; Revised: 6 May 2013; Accepted: 26 August 2013. yra griautinis kapas iš Põlgaste tarand tipo kapo, kuris datuojamas 40 BC – 100 AD laikotarpiu. Gali būti, Raili Allmäe, kad čia yra antrinis palaidojimas, kai buvo sudeginti department of Archaeobiology and Ancient Technology ankstesni nedeginti kaulai. Skirtinguose kapuose ap- Institute of History, Tallinn University tiktų mirusiųjų degintų kaulų spalva skiriasi. Deginant Ryytli 6, 10130, Tallinn, ESTOnIA E-mail: [email protected] lauže, kurio temperatūra siekia 600 – 800 ºC, kaulai tampa pilkos spalvos. Tuo tarpu žmogų deginant tem- peratūroje, viršijančioje 800 ºC, kaulai tampa balti. PASTABOS APIE GELEŽIES Beveik visuose tirtuose kapuose pagal kaulų fragmen- AMŽIAUS LAIKOTARPIO tus galima nustatyti, kad jie buvo ne visai sudeginti, nes KREMACIJĄ ESTIJOJE jie yra tamsiai rudi ir juodi; degintų kaulų kiekiai buvo didesni Maidla (Vakarų Estija) iš akmenų krautuose kapuose ir iš smėlio supiltuose pilkapiuose (Pietryčių RAILI ALLMÄE Estija). Põlgaste tarandų tipo ir Uugla III kapinynuose aptiktų kaulų spalva buvo vienoda. Laidojimo papročiai Rytų ir Vakarų Estijos teritorijose Santrauka skyrėsi. Kūnams deginti skirtų laužaviečių yra aptinkama Vakarų Estijos kapinynuose. Mirusiojo kremacija 1997–2011 m. Estijoje buvo tyrinėti keli laidojimo tikriausiai vyko kapinyno teritorijoje, o kremuoti pa- paminklai su degintiniais kapais. Tyrinėjimų metu laikai išbarstyti šalia. Kaulų trupinimo prieš laidojimą buvo surinkta įvairiausios medžiagos apie degintus papročio neaptikta Pietryčių Estijos pilkapynuose, bet mirusiųjų kaulus. Šio straipsnio tikslas – apibendrinti negalima atmesti prielaidos kad toks paprotys vėly- ir susisteminti duomenis apie Estijos teritorijoje ras- vajame geležies amžiuje čia egzistavo. Daug daugiau tus degintinius palaidojimus, ir pateikti apibendrintus kremuotų kūdikių aptinkama pietrytinėje Estijos teri- duomenis (1–2 pav.; 1–4 lent.). Buvo ištirti degintiniai torijoje. IV–V amžiais datuotame Maidla I kapinyne kaulai, rasti dešimtyje Vakarų ir Pietryčių Estijos lai- taip pat randama daug kremuotų kūdikių. Vėlyvojo ge- dojimo paminklų kapų. Šešiuose kapuose rasti žmonių ležies amžiaus Maidla II kapinyne taip pat aptinkamas kaulai buvo datuoti radioaktyviosios anglies (AMS) kūdikių ir vaikų kaulų sutrupinimo prieš juos palaido- metodu. AMS metodu nustatytos degintinių žmonių jant paprotys, nors vaikų kremacija čia yra labai reta. Iš kaulų datos yra kiek ankstesnės, nei jie buvo datuojami kitų vėlyvojo geležies amžiaus kapinynų yra per mažai anksčiau, kai šie kapai buvo datuoti santykinės chrono- duomenų, todėl išvadų apie buvusius deginimo papro- logijos metodu. Straipsnyje pateikiamas lyginamasis čius daryti negalima. dešimties kapų tyrimas, remiantis minėtomis AMS da- Laidojimo papročių kaita, kurią galima pastebėti tiriant tomis, minimaliu palaidotų individų skaičiumi bei kau- kaulų fragmentacijos etapus, viena vertus, buvo laiki- lu kiekiu ir jų fragmentacija. Vakarinėje Estijoje ištirta nas reiškinys, kaip ir vėlesniuose kapuose, kai sude-

136 ginti kaulai buvo išbarstomi. Pastaruosiuose vidutinis kaulų fragmentacijos lygis yra mažesnis. Norint patikslinti straipsnyje diskutuojamus teiginius, privalu išanalizuoti daugiau chronologiškai skirtingų kremuotų kaulų iš kultūriniu požiūriu skirtingų pamin- klų. BALTICA 19 BALTICA

Vertė Algirdas Girininkas ARCHAEOLOGIA

SOCIETIES Of THE PAST: APPROACHES TO LAndSCAPE

137 138 IV Allmäe, R. (2014). The demography of Iron Age graves in Estonia. Lietuvos Archeologija, 40, 103–120. LIETUVOS ARCHEOLOGIJA. 2014. T. 40, p. 103–120. ISSN 0207-8694

THE DEMOGRAPHY OF IRON AGE GRAVES IN ESTONIA

RAILI ALLMÄE

Cremated and non-cremated human remains from fourteen Estonian Iron Age burial grounds were analysed in order to estimate the number of burials in the graves as well as the age and sex of the individuals in these burials and to model the demographic figures for some Estonian Iron Age communities. Three graves (Rõsna I, Rõsna II, Suure-Rõsna) in SE Estonia and two in W Estonia were suitable for palaeodemographic analyses. Five of the graves were used by communities with 4–9 individuals, which usually corresponds to a single family or household. Rõsna I barrow cemetery in SE Estonia indicated a somewhat larger community of 10–15 individuals, which could have been an extended family, a larger household, or two nuclear families. The estimated crude death rate during the Middle Iron Age at Rõsna varied between 49.2 and 62.1‰ (52.4‰ on average). The estimated crude death rate was somewhat lower (39.0‰) in Late Iron Age Maidla and was extremely high (86‰) in Middle Iron Age Maidla. Keywords: Iron Age, cremations, palaeodemography.

INTRODUCTION ve goods and burial customs of the period and/ or cultural area (Mandel 2003; Allmäe, Maldre The cremated human remains from Estonian 2005; Aun 2005; Allmäe et al. 2007; 2009; Aun Iron Age graves have mainly been studied sin- et al. 2008). Sometimes analyses include the fra- ce the 1990s. The cremated bone analyses have gmentation, colour, and cremation temperature mostly been macroscopic, the main goal being of the bones (Valk, Allmäe 2009; 2010; Allmäe the identification of the number of interred in- 2013). Less attention has been paid to palaeode- dividuals and the determination of their biolo- mographic analyses as there are some require- gical sex and age at death (Kalling 1993; Mägi ments for the material and data, e.g. the whole et al. 1998; Kalman 2000b; Allmäe 2003). These burial site should be excavated, migration should anthropological data have then been combined be excluded, the skeletal material should be with the archaeological data to interpret the gra- complete, accurate data for the burial chronolo-

141 104 RAILI ALLMÄE gy and its start should be available, etc. (Acsádi, new radiocarbon dating methodology (Lanting et Nemeskéri 1970; Alesan et al. 1999). al. 2001) also allows cremated bone material to be The first attempts to analyse the demography dated, which helps in dating cremations where ar- of Estonia’s ancient populations were made on the tefacts are rare and archaeochronological dating basis of the archaeological data (Lang, Ligi 1991; is difficult. This paper presents radiocarbon dates Lang 1996). The authors used the types and num- (AMS method) for cremated human bones from ber of artefacts to calculate the size of the commu- six investigated graves. The goal of these analyses nity, which used the burial site, and made assump- was to determine the timespan when the commu- tions on population density and size based on the nities used these graves. burial grounds in different Estonian districts. The goal of the present study is to systemise the The first palaeodemographic calculations ba- results of the age and sex estimations for the inhu- sed on osteological research from prehistoric gra- mations and cremations in W and SE Estonia from ves were made for the Roman Iron Age Viimsi I the Middle and Late Iron Ages, to calculate some tarand grave in N Estonia (Lang 1993) and for demographic figures, and to draw some conclu- the Pre-Roman Poanse tarand grave in W Estonia sions about ancient burial practices and grave use. (Kalman 2000a). Ken Kalling (1995; 1997) performed the first known palaeodemographic study based on archaeoanthropological material from MATERIALS AND METHODS the Medieval and Early Modern town of Tartu while the author analysed the 13th–14th-century Investigated materials skeletal population in Tartu’s Jaani Church. The author also demographically analysed a skeletal The W Estonian stone graves (Table 1; Fig. 1) sample from the 14th–18th-century Tääksi village were mainly investigated during 1974–1991 by cemetery (S Estonia) (Allmäe 1998); this materi- Archaeologist Mati Mandel (2003), the grave at al contained some 15th–16th-century cremations Keskvere during 2001–2002 (Mandel 2003), and (Соколовский 1990; Allmäe 1998). The author Uugla III stone grave in 2008 (Mandel, Allmäe also demographically analysed skeletal material 2009). The bone material was collected using 2 x from Maidla II (W Estonia) (Allmäe 2006), which 2 m grid squares. The W Estonian graves were all contained 10th–11th-century cremations and 12th– irregular stone constructions while the graves at 13th-century inhumations (Mandel 2003). Leiu Maidla, Kirbla, and Uugla contained areas with Heapost (in 2007) analysed the demography of a thick charcoal layer, which were probably pyre 11th–15th-century Kalmetemägi in Siksälä, SE Es- sites (Mandel 2003). Most of the graves contained tonia. The study concerns mostly inhumations, scattered cremations, although some also had in- but also includes some cremations. humations, for example, from the 5th–7th centuries During 1997–2011, the author analysed several in Lihula, from the 5th–7th/10th–13th centuries in cremation graves in W and SE Estonia, collecting Ehmja, and from the 5th–6th/10th–13th centuries in various descriptive and metric data on cremated the stone graves at Maidla (Mandel 2003). Both bones, including, of course, estimations of the mi- Maidla stone graves were completely excavated, nimum number of buried individuals, the plausi- the second being the biggest ever investigated ar- ble number of interred individuals, the biological chaeologically and osteologically (Allmäe 2003; sex, and the age. The studied material included 2006; Maldre 2003; Mandel 2003). In recent years five completely excavated burial sites, which are new excavations have been conducted in the area, a good source for demographic calculations. The the main purpose being to screen the mounds for

142 THE DEMOGRAPHY OF IRON AGE GRAVES IN ESTONIA 105

Table 1. Investigated Estonian graves

Western Estonia Souhteastern Estonia

Name of the grave I II Kirbla Lihula Ehmja Uugla I Uugla Põlgaste Uugla II Uugla Maidla I Maidla Keskvere Uugla III Uugla Maidla II Maidla Kirikumägi Rõsna-Saare Rõsna-Saare Rõsna-Saare Suure-Rõsna

Type of the grave grave flat-ground flat-ground stone-grave stone-grave stone-grave stone-grave stone-grave stone-grave stone-grave stone-grave sand barrow sand barrow sand barrow tarand

5–7, Period (centuries)1* 5–6 10–13 7–8 5–7 11–13 11–13 11–13 11–13 6–8 7–8 6–8 10–12 3–5 10–13 Number of bone units 268 674 10 144 248 157 30 326 70 159 97 164 69 59 Archaelogically determined nested bone – – – – – – – – – 65 53 41 – – units2** MNI 19 51 1 10 5 4 1 7 1 34 22 26 4 18 PNI 20 74 3 21 10 4 1 11 2 61 46 40 4 23 Cremations 17 42 3 10 5 4 1 11 2 61 46 40 3 23 Inhumations 3 32 – 9 5 – – – – – – – 1 – Subadults (0–15) 8 22 1 4 5 0 0 1 1 21 18 15 1 4 Adults (over 15) 12 52 2 17 5 4 1 10 1 40 28 25 3 19 Males 5 13 – 2 2 – – 2 – 14 10 8 2 12 Females 5 16 – 4 – 1 1 3 1 15 8 8 1 6 Undetermined 2 32 2 11 3 3 0 5 0 11 10 9 1 1

* Dates from Aun 1992; Laul 2001; Mandel 2003; Valk, Allmäe 2010. ** After Aun et al. 2008.

additional archaeological and osteological finds. graves are burial sites with characteristic qua- This has resulted in an interesting discovery of drangular stone enclosures, commonly with N–S a double inhumation below one mound; this is orientations (Jaanits et al. 1982, p.207; Lang 2007, probably a secondary burial or the reburial of the pp.170, 192). The bone material of Põlgaste tarand remains of an adult female and a child (Mandel, grave, all of it cremated, was collected in 1 x 1 m Allmäe 2013). grid squares (Laul 2001, p.42). The cremains were A total of 1927 bone finds from nine Iron Age analysed in 2010–2011. This grave was chosen for W Estonian graves were analysed. comparative analyses for several reasons. The first The SE Estonian osteological material comes was its archaeochronological dating, i.e. the 3rd– from five different graves (Table 1; Fig. 1). Põl- 5th centuries ad (contemporaneous with Maidla gaste tarand grave was excavated by Silvia Laul in I grave). The second was its stone construction, 1970–1973 and dates to the Early Iron Age, i.e. the i.e. tarand grave. It also has other intriguing cha- 3rd–5th centuries (Laul 2001, pp.40–43). Tarand racteristics (Laul 2001, pp.27, 40–41), e.g. the un-

143 106 RAILI ALLMÄE

Fig. 1. Locations of the investigated Estonian Iron Age graves. Map by R. Allmäe. derground Bronze Age cremation burial discove- Mare Aun (Аун 1992) investigated several Long red under it (the charcoal among the bones had Barrow Culture sand-barrow cemeteries in N a 68.2% probability of being from 1260–920 bc) Setumaa from the second half of the 1st millen- and the three sand barrows that lay partially atop nium ad. The present paper analyses three sand- it. Laul (2001, pp.40–43) also detected differences barrow cemeteries (Suure-Rõsna, Rõsna-Saare in the cremated bone material, the bones in the I, and Rõsna-Saare II), which were selected be- grave’s older part being less burnt than those in cause they have been completely excavated. The the later part. A secondary burial custom, i.e. the sand barrow cemeteries in N Setumaa consist of cremation of human skeletal remains, was proba- long, rounded mounds of piled sand that reve- bly observed in the older part (Allmäe 2013). The al various burial customs. The cremated bones complete bone material would have been good were commonly buried in fairly compact assem- source for observing possible changes in the bu- blages or in different containers under and in the rial practices, but unfortunately the osteological mounds (Аун 1992; Aun 2005). The human bone materials collected from Põlgaste tarand grave in material collected from the barrows had all been SE Estonia were only partly available in the depo- cremated (Allmäe, Maldre 2005; Allmäe et al. sitory and came from the grave’s older part. 2007; Allmäe 2013). During the 1970s–1980s Archaeologist During 2003–2004 and 2007, Heiki Valk exca-

144 THE DEMOGRAPHY OF IRON AGE GRAVES IN ESTONIA 107 vated Kirikumägi flat cemetery from the late 10th– hand, in most cases the bone fragments are collec- 12th centuries. Scattered cremation burials were ted using fairly large grid squares: 2 x 2 m. On the discovered under the base. The bones had mainly other hand, there is no way to know how much been cremated, with the exception of some teeth material was carried from the pyre to the burial and temporal bones of a 3–4 year-old child (Valk, site. The material was analysed and the PNI was Allmäe 2010). determined using the criteria described above. A total of 548 bone finds from five Iron Age graves in SE Estonia were osteologically analysed. The SE Estonian sand barrow cemeteries The total material analysed during the study The material from the barrows was also col- consisted of 2475 bone finds. lected as bone assemblages, which were bigger here (except Siksäla flat cemetery). The MNI was Criteria and methods used to estimate the estimated for each barrow in all of the analysed minimum and probable number of individuals sand barrow cemeteries. Many different recurrent bone fragments were recorded, but here also The W Estonian stone graves the pars petrosa of the os temporale, as the most It is known that some parts of a skeleton frequently found element, was the best unit for are more resistant to heat-induced changes and calculating the MNI. Unfortunately the number survive better among cremated bones (Holck of cranial fragments was not counted because it 1997/2008). Thus, the minimum number of indi- initially seemed that the bone assemblages were viduals (MNI) among the examined cremains is well-defined units, i.e. archaeologically distin- very often determined using the recurrent bone guished burials. The collected and deposited units fragment method. Several different bone fra- were named main and additional assemblages. It gments were counted during analyses, but in most was not always possible to decide whether or not cases the pars petrosa of the os temporale was the the additional bone assemblage(s) were part of best unit for estimating the MNI. In some cases, the main one and sometimes it seemed that num- where recurrent fragments did not occur, the in- bered main bone assemblages did not equal one dividuals were distinguished on the basis of biolo- burial. The discrepancy between the MNI and the gical age estimations (e.g., if the bone assemblage archaeologically determined number of burials/ included adult cranial vault fragments as well as assemblages is obvious; the osteologically estima- an unfused long bone epiphysis or deciduous to- ted MNI is usually smaller than the archaeologi- oth crowns, etc.). cally estimated number. There are two reasons for The probable number of individuals (PNI) this. First, the burial criteria are not clearly distin- was estimated on the basis of the fragments (20 or guishable during excavations. It is nearly impos- more) of cranial vault combined with at least one sible to decide whether an assemblage belongs to determined fragment of human skeleton. The cra- an adjacent main assemblage or constitutes a se- nial part of the skeleton was preferred because the parate burial. Second, as always, the percentage of fragments are easily distinguished among crema- the cremains collected from the pyre and buried ted bones and mean that complete corpses were in the barrows remains unknown. probably cremated. The distances between bone Finally, the PNI was determined by taking into units were also taken into account in estimating account the distances of the collected bone units, the PNI, which is an important unit for graves the MNI, the colour of the cremains, the size and with scattered or disturbed cremations (Maidla I– composition of the collected bone unit, and the II, Ehmja, Kirbla, Uugla I–III, Lihula). On the one results of age and sex determinations. Compared

145 108 RAILI ALLMÄE

to the previously published studies (Allmäe, Mal- of the teeth of older individuals become more dre 2005; Allmäe et al. 2007; Aun et al. 2008), this rounded due to the deposition of cementum. Hy- paper uses some PNI and MNI recalculations for percementosis is quite common in older indivi- the sand barrow cemeteries. duals (Acsádi, Nemeskéri 1970; Soames, Southam 1993). Age-related pathologies on cremated bo- The tarand graves and flat cemeteries nes, e.g., osteoarthritis on vertebrae or anywhere In both cases the osteological material was on skeletal elements, were also used. It must be collected using grid squares. The bone assem- emphasized that age and sex determinations made blages in Kirikumägi flat cemetery at Siksälä are for cremated human remains are less reliable than smaller than those in Põlgaste tarand-grave. The those for inhumations. The incompleteness of the MNI and PNI were estimated for both cemeteries. cremated remains due to high bone fragmenta- Unfortunately, the materials represent only part of tion means that only a few skeletal elements are these burial sites: Kirikumägi has been only par- available for determinations. tially excavated and part of the bone material is missing from Põlgaste tarand grave. Demographic estimations The demographic estimations were made Radiocarbon dating using several different methods. Our model assu- Nine samples of cremated human bones from mes that the population is stationary and that the six investigated graves were dated. Cremated bone birth and death rates are equal (growth = 0) be- fragments from Maidla I (Hela-2403) and Maidla cause the Estonian populations under study are all II (Hela-1958) and non-cremated bone (Hela- too small to model a positive or negative natural 1919) from the stone graves, Rõsna-Saare I (Hela- increase. 1959 Hela-1960) and II (Hela-1961) barrow ce- The life table method proposed by G. Acsádi meteries, Suure-Rõsna (Hela-1962) barrow ceme- and J. Nemeskéri (1970) was initially used to esti- 0 tery, and Põlgaste tarand grave (Hela-2404, Hela- mate the life expectancy at birth (e 0). The natural 2405) were radiocarbon dated in 2009 and 2010 data obtained from the skeletal samples was used at the Dating Laboratory of the Finnish Museum for this. of Natural History, University of Helsinki. The 14C The life tables were then corrected in accor- concentration was measured using the AMS met- dance with F. W. Rösing and R. Jankauskas (1997) hod. The results were calibrated using an Intcal09 by increasing the proportion of small children curve (Reimer et al. 2009) and OxCal 4.1 software (0–4 years) in the population to 45% of the total (Bronk Ramsey 2009). skeletal population under study. This means that 45% of population died before they reached the Methods and criteria used for sex age of 5. and age determination J.-P. Bocquet and C. Masset (Bocquet, Masset The sex and age of the individuals were de- 1977; Bocquet-Appel, Masset 1982) established a termined using common osteological standards third model for estimating the demographic figu- (Miles 1963; Workshop 1980; Brothwell 1981; res of past populations. The ratio of subadults to Buikstra, Ubelaker 1994; Bass 2005; Mays 2006). adults (juvenile index) was also calculated for eve- If applicable, other criteria were used to estimate ry population under study: or determine a cremated individual’s age at death: number of children deceased between 5 and cranial vault morphology (Gejvall in Sigvallius 15 / number of adults deceased at 20 and later:

1994) and tooth root morphology, i.e. the roots D5–14/D20+

146 THE DEMOGRAPHY OF IRON AGE GRAVES IN ESTONIA 109

The model was adjusted to overcome the pro- somewhat different from that used in conventio- blem that the number of small children (0–4 ye- nal palaeodemography. Henneberg (1975) com- ars) is often underrepresented in the burials. bined the demographic (mortality structure) and The formula (Bocquet, Masset 1977; Bocquet- biological characteristics of human fertility to Appel, Masset 1982) to estimate the newborn life construct a reproduction model for human palae- expectancy from the juvenile index is: opopulations. In order to estimate a population’s 0 e 0=78.721*log10√1/x–3.384±1.503 reproduction rate, the following definitions and

where x=D5–14/D20+ calculations were used: the potential gross repro-

The newborn life expectancy, crude death rate duction rate (Rpot) or the average number of birt- 0 (1/e 0), and size of the living populations were cal- hs per adult couple during their lifetime, the net culated using all three models. In these models reproduction rate (Ro) or the average number of it was assumed that the population was stationa- adult descendants per adult parent, and the abso- ry and that the birth and death rates were equal lute number of offspring born to an average adult (growth = 0). couple (C). The calculation of the last figure requi-

The size of the living population was calcula- red the hypothetical value Uc (the total number of ted using D. Ubelaker’s formula (1989): births achievable throughout the full reproducti- 0 P=N*e 0/T ve period). The cU closest to reality is about eight where P – population size, (Acsádi, Nemeskéri 1970), but the number could N – number of burials in the cemetery, be lower or higher. In the present study the num- 0 e 0 – life expectancy at birth (in years), ber of offspring born to an average couple (C) was

T – timespan during which the burial site was calculated using the value of Uc=7.45 (Lorimer used (in years). 1954 in Henneberg 1975). The model made an assumption concerning The masculinity index for every population the timespan for the cemetery’s use. In order to was calculated by dividing the number of males model the population size, the 68.2% probability by the number of females in the population. dates (interval) of the cremated bones from each The available data on contemporaneous La- grave were equated with the grave’s usage period tvian and Lithuanian communities were used to (T) with the exception of Maidla II, where 250 discuss the demographic figures of the studied Es- years of grave usage was established by Mandel tonian Iron Age communities (Jankauskas 2002; (2003) and was used in the present calculations. Zariņa 2009). The demographic figures were Reproduction was estimated according to two calculated according to the juvenile/adult ratio different models. The first model estimated the (Boquet-Appel, Masset 1982) with the GRR cali- fertility rate, i.e. the number of female offspring brated in accordance with R. McCaa (1998; 2000). born per woman (gross reproduction rate or GRR), from the juvenile indices of J.-P. Boquet- Appel and C. Masset (1982) and calibrated it in RESULTS accordance with R. McCaa (1998; 2000). It then calculated the total number of offspring per wo- The studied graves are of different sizes and man (total fertility rate or TFR): GRR x 2.05 = from different periods; the composition of the TFR. bone material also varies (Table 1) due to the bu- M. Henneberg (1975) established the second rial practice (cremation, inhumation). In many W model used here to estimate reproduction from Estonian graves (Maidla I and II, Ehmja, Lihula), archaeological human remains. The approach is both cremations and inhumations were found

147 110 RAILI ALLMÄE

(Allmäe 2003; Mandel 2003). Cremation predo- The phenomenon that the radiocarbon dating minated (Аун 1992; Laul 2001; Allmäe 2013) in of cremated bone could yield a somewhat older the studied SE graves, except in Kirikumägi flat age is supported by experimental evidence (Hüls cemetery at Siksälä where the unburnt fragments et al. 2010; Olsen et al. 2013). Hüls and his co- of the skeleton of a 3–4-year old child were found workers (Hüls et al. 2010) found that depending (Valk, Allmäe 2010). In Suure-Rõsna barrow ce- on the cremation temperature, its duration, the metery, some unburnt human vertebrae were composition of the burning atmosphere, and the found in one bone assemblage; this could have composition/age of the fuel, an aging effect of 50– been due to an unsuccessful cremation where the 100 years may be possible. Therefore the possibili- temperature and oxygen flow were insufficient ty cannot be excluded that the relatively old AMS for proper burning of the corpse. The MNI and dates are due to the employed burial techniques. the PNI differ in most cases, the discrepancy, as For example, in the case of the Põlgaste tarand expected, being bigger in the larger graves (Table 1). grave it is highly likely that bare human bones The radiocarbon dates of cremated bones were cremated rather than corpses (Allmäe 2013). from six investigated graves indicate some dis- The biological sex of the individuals was often crepancies compared to the archaeochronological undeterminable, especially in graves with scattered dates (Tables 1, 2). The number of AMS dates is cremations, but masculinity indices were still cal- small, but they help to identify the timespan when culated (Tab 3). In the sand barrows, the index was the grave was in use. According to AMS dates, slightly above 1.0, indicating a balanced sex ratio in Maidla I is from the 4th–5th centuries. The crema- the community. At Maidla II, it was 0.81; a masculi- tions in Maidla II are from the 10th–12th centuries nity index below 1 may signify turbulent times, for as Mandel has already suggested (2003). The ra- example, when men often died away from home. diocarbon dating of one infant inhumation from The proportion of subadults in the graves is this grave showed that in addition to the 12th–13th- 40% or below (Tables 1, 3). The proportion has, century inhumations (Mandel 2003), the infant of course, less importance in small graves (Uugla inhumation(s) could even be from the Middle I–III, Kirbla), where only a few individuals were Ages (Table 2). The sand barrows from SE Estonia buried and/or burned at the site (Allmäe 2003; were radiocarbon dated to the 4th–6th centuries, 2013; Mandel 2003). In partially investigated gra- indicating a somewhat earlier establishment of the ves (Ehmja, Keskvere, Põlgaste, Kirikumägi), the grave compared to the archaeochronological da- age structure can be distorted because only part tes (Tables 1, 2). The dating also confirms the sug- of the osteological material was available for ant- gestion that the barrows were created over a short hropological study. Therefore the proportion of period (Аун 1992; Aun 2005). The flat cemetery at subadults is significant in graves, which have been Siksälä was radiocarbon dated using the charcoal completely excavated: Maidla I–II, Rõsna-Saare found amongst the cremains (Valk, Allmäe 2010), I–II, and Suure-Rõsna. The proportion of suba- which yielded 10th–11th-century dates (Table 1). dults (under 15 years of age) is apparently very The Põlgaste Tarand grave was archaeochronolo- similar for these graves: from 34.4% to 40.0%. The gically dated to the 3rd–5th centuries (Laul 2001, only exception is Maidla II, where the proportion pp.27, 40–41). The radiocarbon dating of crema- of subadults is slightly lower at 29.7%. Even then ted bones from the older part of the grave indica- the proportion of subadults could be overestima- tes a somewhat earlier establishment of the burial ted here, because it is highly probable that the in- site, i.e. in the late Pre-Roman Age or early Roman fant burials or at least some of them are from the Iron Age (Tables 1, 2). Medieval period (Table 2).

148 THE DEMOGRAPHY OF IRON AGE GRAVES IN ESTONIA 111

Table 2. Radiocarbon dating BP C(‰) grave 95.4% 68.2% 13 Lab no. d Location in Bone sample Construction Burial place Burial

adult, left ulna, stone grave, Maidla I proximal, 104/F Hela-2403 -22.4 1675±30 BP 258–427 AD 341–413 AD irregular cremated stone grave, adult, cranial Maidla II 96N–96O Hela-1958 -22.2 1000±30 BP 980–1160 AD 990–1120 AD irregular vault, cremated infant, os stone grave, temporale, Maidla II 75S Hela-1919 -20.7 440±30 BP 1410–1610 AD 1430–1465 AD irregular pars petrosa, uncremated bone set 1, long sand adult, axis, Rõsna-Saare I under barrow 7, Hela-1959 -27.2 1565±35 BP 410–570 AD 430–540 AD barrow cremated cremated subadult, long bone set 3, in the rounded Rõsna-Saare I bone fragment, center of barrow Hela -1960 -26.7 1595±35 BP 390–550 AD 420–540 AD sand barrow cremated 9, cremated bone set 3, in the long sand adult cranial Rõsna-Saare II center of barrow Hela-1961 -24.4 1620±35 BP 340–540 AD 390–540 AD barrow vault, cremated 7, cremated bone set 7, in the rounded sand adult cranial Suure-Rõsna center of barrow Hela-1962 -22.5 1535±35 BP 420–600 AD 430–580 AD barrow vault, cremated 6, cremated tarand adult cranial 40 BC – 122 Põlgaste 8B–8C Hela-2404 -23.9 1959±30 BP 5–74 AD grave vault, cremated AD adult cranial Põlgaste tarand graves 10F Hela-2405 -20.9 1977±30 BP 45 BC – 80AD 20BC – 65AD vault, cremated

The modelled demographic characteristics proposes the lowest newborn life expectancies. for the Estonian graves under study are presented The juvenile indices (D5–14/D20+) for the sand in Table 3. Newborn life expectancy, calculated barrow cemeteries vary from 25 to 32 (27 on ave- 0 using uncorrected life tables (raw data), shows rage) and e 0 between 16.3 and 20.3 years (19.0 the highest values for the graves, signifying a years on average), indicating a mortality betwe- relatively low mortality. The data, corrected in en 49.2 and 62.1‰ (52.4‰ on average). These accordance with Rösing and Jankauskas (1997) mortality values are much higher compared to by increasing the proportion of infants and chil- those obtained from the uncorrected life tables dren (0–4 years) to 45% of the total number of in- (31–37‰) and slightly higher compared to the dividuals under study, yields lower and probably mortality obtained from the corrected life tables more reliable values for newborn life expectancy (46–57‰). (Table 3). The stone graves of Maidla show greater va- The model based on the juvenile ratio (Bo- riability in the juvenile indices and newborn life cquet, Masset 1977; Bocquet-Appel, Masset 1982) expectancy (Table 3). Maidla I shows a very high

149 112 RAILI ALLMÄE

Table 3. Demographic data from the Estonian graves Maidla I Maidla Maidla II Maidla Suure-Rõsna Rõsna-Saare I Rõsna-Saare Rõsna-Saare II Rõsna-Saare Three barrow Name/type of the burial place cemeteries, Type of the grave summarised stone-grave stone-grave sand barrow sand barrow sand barrow

AMS dates 68.2% (range) 341–413 AD 990–1120 AD 420–540 AD 390–540 AD 430–580 AD 390–580 AD AMS dates 95.4% (range) 258–427 AD 980–1160 AD 390–570 AD 340–540 AD 420–600 AD 340–600 AD Number of burials 20 74 61 46 40 147 Males 5 13 14 10 8 32 Females 5 16 15 8 8 31 Index of masculinisation (%) 1.0 0.81 0.87 1.25 1 1.03 Undetermined sex 2 23 11 10 9 30 Proportion of adults (%) 60 70.30 65.60 60.90 62.50 63.27 Proportin of subadults (%) 40 29.70 34.40 39.10 37.50 36.73 T = usage of grave (years) 72 2501* 120 150 150 190 Life tables raw data after Acsádi, Nemeskéri 1970 0 e 0 newborn life expectancy 26.9 28.2 28.42 27.25 32.02 29.03 Crude death = crude birth rate 0.037 0.035 0.035 0.037 0.031 0.034 Corrected life tables** 0 e 0 newborn life expectancy 17.5 19.56 19.56 19.89 21.72 20.6 Crude death = crude birth rate 0.057 0.051 0.051 0.050 0.046 0.049 After Boquet, Masset 1982 Juvenility index 5–14/20+ 0.42 0.18 0.25 0.25 0.32 0.27 0 e 0 newborn life expectancy 11.6 25.7 20.3 20.3 16.3 19.0 Crude death = crude birth rate 0.086 0.039 0.049 0.049 0.061 0.053 GRR gross reproduction rate 5.5 2.5 3.1 3.1 3.9 3.2 TFR total fertility rate 11.3 5.1 6.4 6.4 8.0 6.6 Reproduction after Henneberg 1975

Rpot 0.688 0.668 0.675 0.680 0.776 0.709

R0 (UC=7.45) 1.52 1.80 1.53 1.54 1.76 1.61 C – average number of births 5.0 5.1 5.0 5.1 5.8 5.3

Population size*** 3.2–7.5 7.6–8.7 10.3–14.8 6.2–8.5 4.3–8.7 14.7–22.8

* From Mandel 2003. ???????????? **After Rösing, Jankauskas 1997. ***After Ubelaker 1989.

150 THE DEMOGRAPHY OF IRON AGE GRAVES IN ESTONIA 113 crude death rate (88‰) and an extremely low DISCUSSION 0 newborn life expectancy (e 0=11.4); under these conditions a community is hardly sustainable. The earlier demographic studies of the Esto- At Maidla II, the mortality is the lowest (39‰) nian prehistoric and historic periods have been 0 and the life expectancy at birth (e 0=25.7) is the based on different assumptions and modelled highest compared to the other studied commu- using various methods. The first demographic nities. Bear in mind that Maidla I is a very small model to calculate community sizes for the pre- grave (only 20 burials) and at Maidla II both historic period was based on the number of arte- infants and older children (over 5 years of age) facts in tarand graves and a presumed mortality could be underrepresented, only two cremations rate of 40‰. The suggested average estimated size of subadults having been detected in the grave of the community that used one tarand grave was (Allmäe 2003). The GRR and TFR obtained from 5–9 individuals in NE, SE and Central Estonia and the juvenile indices are also presented in Table 3–4 individuals in NW Estonia (Lang, Ligi 1991, 3. According to this model, the women at Late pp.224–225). Later on Valter Lang (1996, p.375) Iron Age Maidla are characterised by the lowest corrected the aforementioned numbers to 7–13 number of offspring, the Middle Iron Age wo- and 3–8 individuals, respectively. A community men in Maidla by the highest. The Middle Iron size of 8–10 individuals was obtained from the Age women at Rõsna gave birth to 6.6 children skeletal remains from the Roman Iron Age Viim- on average. si I tarand grave in N Estonia (Lang 1993, p.56), The number of children born per woman which corresponds well with the artefact-based was also calculated using the model proposed by calculations. M. Henneberg (1975). The total number of birt- The osteological research of skeletal material hs per woman or adult couple is from 5.0 to 5.8 from the Pre-Roman Iron Age tarand graves at in the communities under study. The number of Poanse (W Estonia) indicated a community of offspring varies from 5.0 to 5.8 for the SE barrow 4–6 individuals depending on the period of the cemeteries and between 5.0 and 5.1 for the W grave’s use (Kalman 2000a; Lang 2007, p.224).

Estonian stone-graves (Table 3). Ro or the net re- The community at Tandemägi (Võhma, N Esto- production rate shows the replaceability of gene- nia) consisted of 6 individuals (Lang 2000, p.206). rations; the Ro>1 for all of the observed commu- Lang (2007, pp.224–225) pointed out that a single nities suggests a positive increase. family used the tarand graves for centuries and The living population sizes calculated using that custom of burying only nuclear family mem- different life expectancies indicate that in most bers and sometimes only some of them in stone cases one household or family used the graves graves was practiced during the Pre-Roman and under study. The model also suggest that Rõsna- Roman Iron Ages. Saare I cemetery was used by a somewhat larger The first stone grave of Maidla in W Estonia household or extended family or two nuclear fa- was most likely established in the late Roman Iron milies (Table 3). The calculated living population Age or early Middle Iron Age. At most, a single fa- size is sensitive to the estimated period of the mily used the grave; the crude death rate was very 0 grave’s use; as the timespan decreases, the calcu- high (e 0=11.6) and the community was probably lated population size will increase in these mo- not sustainable. The same pattern characterises dels (Table 3). the Pre-Roman Iron Age tarand graves at Poanse

151 114 RAILI ALLMÄE

(Kalman 2000a; Mandel 2000). The newborn life nes before burial in Late Iron Age W Estonia is expectancy was extremely low (10.8 and 14.4 ye- probable (Allmäe 2013); in this case, the fragile ars), thus the crude death rates are very high and cremains of children become invisible in graves the calculated community sizes very small (3.5 pe- (Sigvallius 1994, p.32; Holck 1997). Insufficient ople at Maidla I, 2.0 at Poanse I, and 3.4 at Poanse excavation techniques are less plausible since the II) The period these graves were used is 72, 250, number of subadults in Maidla I is representative. and 150 years, respectively. The timespan assump- In respect to community size, the results of tions are probably wrong and the graves were pos- the osteological analyses of the cremains from the sibly used for a shorter period. Population size Middle Iron Age barrow cemeteries at Rõsna (SE estimation is highly dependent on the period of Estonia) show the same pattern of one family or a grave’s use and the estimated life expectancy at household using one barrow cemetery. The com- birth. Our idea that graves, especially tarand gra- munity size assumptions based on the archeologi- ves, were used for several centuries could be so- cally determined number of burials (Rõsna-Saare mewhat overestimated. For example, if the use pe- I and Rõsna-Saare II) and a mortality rate of 40‰ riod for Poanse I is reduced from 250 to 100 years yield similar results (Лиги 1989; Lang, Ligi 1991, and for Poanse II from 150 to 50 and a newborn p.227). Similar results have been obtained from life expectancy derived from the juvenile ratio Iron Age East Lithuania: communities of 5–15 in- 0 0 (e 0=10.8; e 0=14.4 years) is used, another reality dividuals usually buried their dead in one barrow is seen. The community size for Poanse I tarand cemetery, the number of individuals correspon- grave is then 5 and for Poanse II 10. ding to a group of people the size of the average During the Late Iron Age a community of nuclear family over several generations (Kurila 7–10 individuals, probably a single family or 2009). The newborn life expectancy at Rõsna du- household, used Maidla II in W Estonia (Allmäe ring the Middle Iron Age was 19.0 years, the crude 2006). The estimated family size is in accordan- death rate was 49–53‰ on average, and women ce with the earlier results of various authors on gave birth to 5.3–6.6 children (Table 3), indicating the average Estonian family size in the 13th cen- relatively unfavourable living conditions compa- tury and later (Blumfeldt 1937; Ligi 1961; Tarvel red to Late Iron Age Maidla. 1972; Palli 1996). The present study also suggests The present study shows that the demographic that a single family or household (7–9 individu- figures depend on the selected model or sample. als) probably had its own burial ground at Maidla The overall variability in the proportion of adults during the 10th–13th centuries. The newborn life and subadults in the graves is not striking, but expectancy was 25.7 years, the crude death rate when the skeletal samples are very small, any mi- was 39‰, and women gave birth to 5.1 children nor change in the proportions of the age groups has on average in Maidla during the 10th–13th centu- a significant impact on the demographic figures. ries. The newborn life expectancy shows low child The next step looks more closely at two most mortality and favourable living conditions or un- representative samples from Estonia: the summa- derrepresentation of the subadult (older than 5 rised sample of SE Estonian sand barrows from years of age) burials in the grave. There are many the Middle Iron Age at Rõsna (Rõsna I–II, Suure- possible reasons for child underenumeration Rõsna) and the W Estonian stone grave at Mai- in the graves: segregation in the subadult burial dla II from the Late Iron Age. The demographic practice is plausible (Allmäe 2010), the crushing figures calculated for these samples and for some of the cremains before burial, and the excavation Latvian (Zariņa 2009) and Lithuanian (Jankaus- techniques. For example the crushing of burnt bo- kas 2002) archaeological skeletal materials are

152 THE DEMOGRAPHY OF IRON AGE GRAVES IN ESTONIA 115

Table 4. Comparative demographic data

Southeastern Estonian Lithuania * Grave Rõsna (summarised) Marvelė Marvelė Plinkaigilis Date (years) 390–580 AD 300–450 AD 450–600 AD 450–600 AD Usage of grave (years) 190 150 150 150 Number of burials 147 179 223 334 Juvenility index 0.27 0.32 0.27 0.20 Newborn life expectancy 19.1 16.3 18.7 24.0 Crude death rate 0.052 0.061 0.053 0.042 GRR** 3.2 3.8 3.2 2.6 TFR 6.6 7.8 6.6 5.3

Western Estonia Latvia *** Grave Maidla II Leijasbitēni Čunkāni-Dreņģeri Laukskola Date (years/centuries) 980–1230 AD 7th–10th 8th–11th 10th–13th Usage of grave (years) 250 300 300 275 Number of burials 74 188 233 239 Juvenility index 0.18 0.21 0.11 0.28 Newborn life expectancy 25.7 23.3 34.2 18.3 Crude death rate 0.0390 0.0430 0.0293 0.0546 GRR** 2.5 2.6 1.9 3.4 TFR 5.1 5.3 3.9 6.97

* Data from Jankauskas 2002. ** Calibrated from McCaa 1998; 2000. *** Data from Zariņa 2009. presented in Table 4. The juvenile index is lowest tery from the 10th–13th centuries. The dramatic in Marvelė during 300–450, showing the highest fluctuation in population size there due to econo- mortality in the studied communities. Jankauskas mic and political processes has been thoroughly (2002) has suggested some kind of crisis, which analysed by G. Zariņa (2009, pp.180–184). The led to a dramatic decrease in the Marvelė com- collective influence of these processes is reflected munity during this period, but in the following in the demographic figures of the summarised Middle Iron Age period the situation improved. sample. Rõsna in S Estonia shows juvenile indices compa- The number of offspring (TFR values), as rable to those of Middle Iron Age Marvelė, both expected, increases with the mortality rate be- indicating less suitable living conditions compa- cause higher infant mortality reduces the birth red to the cemetery in Plinkaigalis (Lithuania). A intervals. When a nursing child dies, the natural somewhat better demographic situation is indi- sterility due to the lactation period is interrupted cated for the Estonian and Latvian communities and conception occurs earlier (Wood 1990). The somewhat later during the Middle Iron Age and human reproductive strategies may change due to Late Iron Age. The exception is Laukskola ceme- the influence or interaction of several economic,

153 116 RAILI ALLMÄE social, and biological factors. Fertility may increa- Allmäe, R., Maldre, L., 2005. Rõsna-Saare I kääbas- se, e.g., during wars, famine and drought periods, kalmistu – esialgseid osteoloogilisi andmeid. In: Tamla, even though the mortality rate is high at the same Ü., ed. ST, 3. Uurimusi Setumaa loodusest, ajaloost ja time. The phenomenon is observable in Table 4: folkloristikast. Tallinn: Ajaloo Instituut, 121–137. low newborn life expectancy is related to a higher Aun, M., 2005. Pikk-kääbaste ehitusest. In: Tamla, number of offspring. Ü., ed. ST, 3. Uurimusi Setumaa loodusest, ajaloost ja folkloristikast. Tallinn: Ajaloo Instituut, 97–120. Acknowledgements Aun, M., Allmäe, R., Maldre, L., 2008. Pikk- The study was undertaken within the fra- kääbaste tähendusest (Rõsna küla kääbaskalmistute mework of a target research project funded by the materjali põhjal). In: Tamla, Ü., ed. ST, 4. Uurimusi Estonian Government (SF0130012s08). Setumaa loodusest, ajaloost ja folkloristikast. Tallin- na Ülikooli Ajaloo Instituut, 269–290. Bass, W.M., 2005. Human Osteology: A Labora- REFERENCES tory and Field Manual, 5th ed. Missouri Archaeolo- Acsádi, G.Y., Nemeskéri, J., 1970. History of Hu- gical Society. man Life Span and Mortality. Budapest: Akadémiai Blumfeldt, E., 1937. Eesti majandusajalugu, I. Kiadó. Tartu: Akadeemiline Koperatiiv. Alesan, A., Malgosa, A., Simó, C., 1999. Looking Bocquet, J.-P., Masset, C., 1977. Estimateurs en into the demography of an Iron Age population in Paléodémographie. L‘Homme, XVIII, 65–90. the Western Mediterranean. I. Mortality. American Bocquet-Appel, J.-P., Masset, C., 1982. Farewell Journal of Physical Anthropology, 110, 285–301. to Paleodemography. JHE, 11 (4), 321–333. Allmäe, R., 1998. Tääksi 14.–18. sajandi popu- Bronk Ramsey, C., 2009. Bayesian analysis of ra- latsiooni demograafiline analüüs ja kehapikkuse re- diocarbon dates. Radiocarbon, 51 (1), 337–360. konstrueerimine. MT, 5, 163–187. Brothwell, D.R., 1981. Digging up bones: The exca- Allmäe, R., 2003. Läänemaa 5.–13. sajandi kal- vation, treatment and study of human skeletal remains, mete antropoloogiline aines. In: Mandel, M., Lääne- 3rd ed. London: Trustees of the British Museum. maa 5.–13. sajandi kalmed, Appendix 1. EAMTAA, Buikstra, J.E., Ubelaker, D.H., eds., 1994. Stan- 5. Tallinn: Eesti Ajaaloo Muuseum, 243–262. dards for Data Collection from Human Skeletal Re- Allmäe, R., 2006. Grave 2 of Maidla – the burial mains. Proceedings of a Seminar at the Field Museum site of a single family. EJA, 10 (1), 3–23. of Natural History Organized by Jonathan Haas Allmäe, R., 2010. Some remarks on Kaseküla (=Arkansas Archeological Survey Research Series, stone-cist grave, Läänemaa, Estonia. Fennoscandia 44). Fayetteville: Arkansas Archeological Survey. Archaeologica, XXVII, 45–52. Heapost, L., 2007. The cemetery of Siksälä: oste- Allmäe, R., 2013. Observations on Estonian Iron ological and paleodemographical analysis. In: Laul, Age cremations. Archaeologia Baltica, 19, 31–47. S., Valk, H., eds. Community at the Frontiers. Iron Allmäe, R., Aun, M., Maldre, L., 2007. Crema- Age and Medieval. Tartu University, 213–236. tions of the Culture of Long Barrows in Northern Henneberg, M., 1975. Notes on the reproduction Setumaa in the Second Half of the First Millennium. possibilities of human prehistorical populations. Pr- Preliminary Results. Humanbiologia Budapestinen- zegląd Antropologiczny, XLI, 75–89. sis, 30, 113–122. Holck, P., 1997. Why are small children so seldom Allmäe, R., Aun, M., Maldre, L., 2009. Rõsna- found in cremations? In: Smits, E., Iregren, E., Drusi- Saarõ I kääbaskalmistu. In: Aun, M., ed. Setomaa 2. ni, A.G., eds. Cremation Studies in Archaeology (Pro- Vanem ajalugu muinasajast kuni 1920. aastani. Tar- ceedings of the Symposium Amsterdam 26.–27. October tu: Eesti Rahva Muuseum, 88–94. 1995, Amsterdam). Saonora: Logos Edizioni, 33–38.

154 THE DEMOGRAPHY OF IRON AGE GRAVES IN ESTONIA 117

Holck, P., 1997/2008. Cremated Bones. A Medi- Lang, V., 2000. Keskusest ääremaaks. Viljelusma- cal-Anthropological Study of an Archaeological Mate- jandusliku asustuse kujunemine ja areng Vihasoo- rial on Cremation Burials, 3rd ed. University of Oslo. Palmse piirkonnas Virumaal. Tallinn: Ajaloo Instituut. Hüls, C.M., Nadeau, M.J., Grootes, P.M., Erlen- Lang, V., 2007. The Bronze and Early Iron Ages in keuser, H., Andersen, N., 2010. Experimental study Estonia (=Estonian Archaeology, 3). Tartu University on the origin of cremated bone apatite carbon. Ra- Press. diocarbon, 52 (2), 587–599. Lang, V., Ligi, P., 1991. Muistsed kalmed ajaloo- Jaanits, L., Laul, S., Lõugas, V., Tõnisson, E., lise demograafia allikana. MT, 1, 216–238. 1982. Eesti esiajalugu. Tallinn: Eesti Raamat. Lanting, J.N., Aerts-Bijma, A.T., van der Plicht, Jankauskas, R., 2002. Anthropology of the J., 2001. Dating of cremated bones. Radiocarbon, 43, Iron Age inhabitants of Lithuania. In: Bennike, P., 249–254. Bodszár, É.B., Susanne, C., eds. Ecological Aspects of Laul, S., 2001. Rauaaja kultuuri kujunemine Past Human Settlements in Europe. Biennal Books of Eesti kaguosas. Tallinn: Ajaloo Instituut – Õpetatud EAA, 2. Budapest: Eötvös University Press, 129–142. Eesti Selts. Kalling, K., 1993. Viimsi kalmete luuainese an- Ligi, H., 1961. Eesti talurahva olukord ja klassi- tropoloogiline analüüs. In: Lang, V., ed. Kaks tarand- võitlus Liivi sõja algul (1558–1561). Tallinn: ENSV kalmet Viimsis, Jõelähtme kihelkonnas, Appendix 1. Teaduste Akadeemia. Tallinn: ETAA Instituut, 67–69. Mägi, M., Allmäe, R., Maldre, L., 1998. Viking Kalling, K., 1995. Paleoantropoloogilisi an- Age graveyard at Piila, Saaremaa. AVE 1997, 99–116. dmeid Tartu Jaani kiriku kalmistu 13.–14. sajandi Maldre, L., 2003. Läänemaa kivikalmete arhe- matuste kohta. In: Valk, H., ed. TÜAKT, 8. Tartu ozooloogiline aines. In: Mandel, M. Läänemaa 5.– arheoloogiast ja vanemast ehitusloost. Tartu Ülikooli 13. sajandi kalmed. Appendix 2. EAMTAA, 5. Tal- Arheoloogia kabinet, 47–57. linn: Eesti Ajaloo Muuseum, 263–286. Kalling, K., 1997. Uusi paleoantropoloogilisi an- Mandel, M., 2000. Poanse tarandkalmed. EAM- dmeid Tartu Jaani kiriku kalmistu kohta. In: Valk, TAA, 2, 89–112. H., ed. TÜAKT, 9. Arheoloogilisi uurimusi, I. Tartu Mandel, M., 2003. Läänemaa 5.–13. sajandi kal- Ülikooli Arheoloogia kabinet, 54–70. med (=EAMTAA, 5). Kalman, J., 2000a. Skeletal analysis of the graves Mandel, M., Allmäe, R., 2009. Ergebnisse Der of Kaseküla, Poanse I and Poanse II. EAMTAA, 2, Archäologischen Ausgrabungen. In: Uugla. AVE 17–40. 2008, 114–123. Kalman, J., 2000b. Uusküla II skeletal analysis. Mandel, M., Allmäe, R., 2013. Forschungs- und MT, 7, 437–440. Herrichtungsarbeiten im Gräberfeld zu Maidla. Kurila, L.V., 2009. The Social Organisation in AVE 2012, 281–288. East Lithuania in the 3rd–12th Centuries (on the Ba- Mays, S., 2006. The archaeology of human bones. sis of the Mortuary Record) (Summary of doctoral London, New York: Routledge. dissertation). Vilnius University. Available from: McCaa, R., 1998. Calibrating paleodemography: http://vddb.library.lt/fedora/get/LT-eLABa-0001:E. the uniformitarian challenge turned. American As- 02~2009~D_20091109_151308-59594/DS.005.1.01. sociation of Physical Anthropology Annual Meeting, ETD [Accessed 23 November 2013]. April 2, 1998, Salt Lake City. Available from: http:// Lang, V., 1993. Kaks tarandkalmet Viimsis Jõeläht- www.hist.umn.edu/~rmccaa/paleo98/index0.htm me kihelkonnas. Tallinn: ETAA. [Accessed 12 November 2013]. Lang, V., 1996. Muistne Rävala. Muistised, kronolo- McCaa, R., 2000. Paleodemography of the Ame- ogia ja maaviljelusliku asustuse kujunemine Loode-Ees- ricas. In: Steckel, R.H, Rose, J.C., eds. The Backbone tis, eriti Pirita jõe alamjooksu piirkonnas, I–II (=MT, 4). of History: Health and Nutrition in the Western He-

155 118 RAILI ALLMÄE

misphere. New York: Cambridge University Press, Valk, H., Allmäe, R., 2010. Kirikumägi at Siksälä: 94–126. evidence of a new grave form of South-Eastern Esto- Miles, A.E.W., 1963. The dentition in the as- nia. EJA, 14, 40–55. sessment of individual age in skeletal material. In: Wood, J.W., 1990. Fertility in anthropological Brothwell, D.R., ed. Dental Anthropology. New York: populations. Annual Review of Anthropology, 19, Pergamon Press, 191–209. 211–242. Olsen, J., Heinemeier, J., Hornstrup, K.M., Ben- Workshop, 1980. Workshop of European Ant- nike, P., Thrane, H., 2013. ‘Old wood’ effect in radio- hropologists. Recommendations for Age and Sex carbon dating of prehistoric cremated bones? Jour- Diagnoses of Skeletons. JHE, 9, 517–549. nal of Archaeological Science, 40, 30–34. Zariņa, G., 2009. Latvijas iedzīvotāju paleodemo- Palli, H., 1996. Eesti rahvastiku ajalugu aastani grāfija: 7. g.t. pr. Kr. – 1800. g. Rīga: Latvijas Vēstures 1712 (=Academia, 6). Tallinn: ETAK. institūta apgāds. Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Аун, М., 1992. Археологичеческие памятники Beck, J.W., Blackwell, P.G., Bronk Ramsey, C., Buck, второй половины 1-го тысячелетия н.э. в Юго- C.E., Burr, G.S., Edwards, R.L., Friedrich, M., Grootes, Восточной Эстонии. Таллинн: Олион. P.M., Guilderson, T.P., Hajdas, I., Heaton, T.J., Hogg, Лиги, П., 1989. Культура длинных курганов A.G., Hughen, K.A., Kaiser, K.F., Kromer, B., McCor- в свете данных палеодемографии. ETATÜ, 38, mac, F.G., Manning, S.W., Reimer, R.W., Richards, 316–317. D.A., Southon, J.R., Talamo, S., Turney, C.S.M., van der Соколовский, В., 1990. Средневековое дере- Plicht, J., Weyhenmeyer, C.E., 2009. IntCal09 and Ma- венс кое кладбище в Тяекси. ETATÜ, 4, 414–422. rine09 Radiocarbon Age Calibration Curves, 0–50,000 Years cal BP. Radiocarbon, 51, 1111–1150. ABBREVIATIONS Rösing, F.W., Jankauskas, R., 1997. Infant deficit in pre-modern burial sites. In: Thetloff, M., ed. The AVE – Arheoloogilised välitööd Eestis = Ar- 8th Tartu International Anthropological Conference. chaeological fieldwork in Estonia 12–16 October, Tartu, Estonia. Dedicated to the 100th EAMTAA – Eesti Ajaloomuuseum. Töid ajaloo Birth Anniversary of Prof. Juhan Aul. University of alalt Tartu, Centre for Physical Anthropology, 50–52. EJA – Estonian Journal of Archaeology Sigvallius, B., 1994. Funeral Pyres. Iron Age Cre- ETAA – Eesti Teaduste Akadeemia Ajaloo Ins- mations in North Spånga (=Theses and Papers in tituut Osteology, I). Stockholm University, Department of ETATÜ – Eesti Teaduste Akadeemia Toimeti- Archaeology, Osteological Research Laboratory. sed, Ühiskonnateadused Soames, J.V., Southam, J.C., 1993. Oral Patholo- JHE – Journal of Human Evolution gy, 2nd ed. Oxford University Press. MT – Muinasaja Teadus Tarvel, E., 1972. Adramaa. Eesti talurahva maa- ST – Setumaa kogumik = Setumaa symposium kasutuse ja maksustuse alused 13.–19. sajandil. Tal- = Сетумааский сборник. linn: Eesti Teaduste Akadeemia Ajaloo Instituut. TÜAKT – Tartu Ülikooli Arheoloogia Kabineti Ubelaker, D.H., 1989. Human Skeletal Remains: Toimetised Excavation, Analysis, Interpretation, 2nd ed. (=Ma- GRR – gross reproduction rate, i.e. the number nuals on Archeology, 2). Washington: Taraxacum. of female offspring born per woman Valk, H., Allmäe, R., 2009. Põletusmatused MNI – minimum number of individuals Siksälä Kerigumäel. In: Aun, M., ed. Setomaa 2. Va- PNI – probable number of individuals nem ajalugu muinasajast kuni 1920. aastani. Tartu: TFR – total fertility rate, i.e. the total number of Eesti Rahva Muuseum, 387. offspring per woman

156 KOKKUVÕTE

Aastatel 1997–2011 uuriti Eesti rauaaegseid laiba- ja põletusmatustega kalmeid ja ühte võimalikku alternatiivset imikute matusepaika. Vaatluse all oli arheoantropoloogiline materjal viieteiskümnest Läänemaa ja Kagu-Eesti matusepaigast. Läänemaalt uuriti kümmet matusepaika (2008 luukogumit) ja Kagu- Eestist viit matusepaika (548 luukogumit). Uurimisperioodil koguti põlenud luude ja matusepaikade kohta erinevaid kirjeldavaid ja meetrilisi tunnuseid. Käesoleva uurimistöö üks eesmärk oli nende tunnuste süstematiseerimine ja interpreteerimine. Teine eesmärk oli hinnata matusepaiku kasutanud kogukondade suurust ja koosseisu. Kolmas eesmärk oli leida võimalikke ajalisi, piirkondlikke ja kultuurilisi erinevusi matusekombestikus keskmisel ja nooremal rauaajal. Seitsme kalme kasutusaja täpsustamiseks kasutati põlenud ja põletamata luude dateerimist radiosüsinikmeetodil (AMS meetod). Põlenud luudest saadud dateeringud erinevad varasematest arheokronoloogilistest dateeringutest, näidates, et kalmeid võidi hakati kasutama mõnevõrra varem kui seni arvatud. Hüpoteesi kinnitamiseks on siiski kindlasti vaja suuremaid radissüsinikdateeringute seeriaid. Kümne kalme võrdlev analüüs viidi läbi radiosüsinikdateeringute, määratatud luukogumite osakaalu ja luude fragmenteerumise põhjal. Määratud luukogumite osakaal on 1. sajandisse eKr – 8. sajandisse pKr dateeritud kalmetes suurem võrreldes 10.–13. sajandi pKr kalmetega, samuti näitavad perioodi 1. sajand eKr – 8. sajand pKr dateeritud kalmed väiksemat luude fragmenteerumise astet. Koljufragmentide ja pikkade toruluude mõõtmed arheoloogilises materjalis kahanevad ajas, see tähendab, et varasemates kalmetes, kus peamiselt esinevad kompaktsemad luukogumid, on luude fragmenteerumine väiksem. Hilisemates kalmetes (10.–13. sajand pKr), kus esineb peamiselt puistepõletusmatus, on põlenud luude fragmenteerumine suurem. Peaegu kõigi põlenud luumaterjalidele puhul on jälgitavad murdumismustrid, mis tekivad juhul kui kremeeritakse laipu. Esineb ka üks erand, Põlgaste tarandkalme (AMS dateering: 1. sajand eKr – 2. sajand pKr), kus tõenäoliselt on tegemist sekundaarse matusega, mille puhul on võimalik, et inimsäilmed põletati alles pärast laiba kõdunemist. Põlenud luukogumite värvus kalmetes varieerub, need on valdavalt valged või heledad, kohati esinevad hallid, sinakad ja helepruunid toonid. Laipade põletustemperatuur on olnud vähemalt 600–800 ºC, tõenäoliselt kõrgemgi. Laipade põletamisel tuleriidal ei põle aga kõik luud ühtlaselt, mistõttu leidus kõigi kalmete materjalis mittetäielikult põlenud luid, mida iseloomustab mustjate ja tumepruunide luukatkete esinemine. Ebatäielikult põlenud luude osakaal oli suurem Maidla kivikalmetes ja Rõsna-Saare kääbaskalmistutes. Erinevustele matusekombestikus võib viidata luude sekundaarne värv, mida põhjustavad erinevused matuseks ettevalmistatava kogumi komplekteerimisel. Kõige paremini olid jälgitavad erisused Kagu-Eesti liivakääbastes, kus osadel juhtudel olid luud tuleasemelt

157 kogutud puhtalt, see tähendab, et kogumis olid luud ilma tahmakihita. Samal ajal oli osa luukogumeid maetud koos tuleaseme jäänustega, mistõttu maetud luud olid tahmased. Seega võib öelda, et samaaegselt esines kaks erinevat viisi käsitseda inimsäilmeid pärast tuleriida kustumist ja ettevalmistamisel matuseks. Läänemaa kivikalmetes toimus laipade põletamine sageli kalme alal ning luud (või osa neist) puistati laiali põletuskoha lähiümbruses, mistõttu leidub seal sageli samuti tahmaseid luid. Matusekombestik Läänemaal ja Kagu-Eestis on olnud erinev ka laibamatuste ja põletusmatuste sisalduse osas. Kui Läänemaa kalmetes esines sageli koos nii laiba- kui ka põletusmatuseid, siis uuritud Kagu-Eesti matusepaikadest leiti ainult põletatud inimsäilmeid. Läänemaa kalmetes leiti sageli tahmaseid pinnaselaike, milles sisaldus väikesi luude ja esemete katkeid, tõenäoliselt põletati laipu kalme alal ning põlenud luud maeti või puistati põletuskoha lähedusse. Uuritud Kagu-Eesti matusepaikadel põletuskohti ei leitud, tõenäoliselt toimus põletamine matusepaikadest eemal, võimalik, et asula lähedal, kus krematsiooniprotsessi oli lihtsam kontrolli all hoida. Kagu-Eesti kääbaskalmistutes luude tahtlikku purustamist enne matust ei tuvastatud, kuid seda ei saa välistada noorema rauaaja kalmete (10.–13. sajand) puhul. Laste ja imikute põletusmatuseid esines rohkem (4.)6.–8. sajandi Kagu-Eesti kääbaskalmistutes, kuid laste matuste osakaal oli suur ka (4.)5.–6. sajandi Maidla I kivikalmes. Maidla noorema rauaaja kivikalmes (10.– 13. sajand) oli vaid üksikuid laste põletusmatuseid, imikute matuseid ei leitud ühtegi. Võimalik, et lapsi ja imikuid maeti mujale või on nende haprad säilmed erinevate protsesside käigus hävinud (sh matusekombestik, protsessid pinnases ja tegevused arheoloogiliste uuringute käigus). Alternatiivse matusepaigana noorema rauaaja imikutele tuvastati näiteks Kaseküla noorema pronksiaja kivikirstkalmed. Teiste uuritud noorema rauaaja kalmete kohta ei saa laste matuste osas olulisi järeldusi teha nende väiksuse või osalise uurituse tõttu. Muutused ja erinevused matusekombestikus, mida me adume läbi luumaterjali fragmentatsiooni ja määratavate luukogumite arvu, on ajalised; hilisemates puistepõletusmatusetga kalmistutes on luukatkete keskmine ja ka mediaanpikkus väiksemad. Ajalised ja kultuurilised erinevused ilmnevad aga ka laste ja imikute matusekombestikus. Täpsemalt, laste ja imikute põletusmatused on küllaltki hästi esindatud pikk-kääbaste kultuuri matusepaikades (4.)6.–8. sajandil ning Maidla (4.)5.–6. sajandi kivikalmes. Leitud suundumuste ja erisuste kinnitamiseks matusekombestikus on vajalik Eesti põletusmatuste osteoloogilise materjali laialdasem uurimine, kus nii kultuurline kui ajaline komponent varieeruks. Paleodemograafiliseks analüüsiks olid sobivad viis matusepaika: Kagu-Eestist Rõsna-Saare I ja II ning Suure-Rõsna kääbaskalmistud keskmisest rauaajast; Läänemaalt Maidla I kivikalme keskmisest rauaajast ja Maidla II kivikalme nooremast rauaajast. Kalmete kasutusaeg täpsustati radiosüsinikdateeringute (AMS) abil, demograafilise mudeli koostamiseks määratleti üksikute kääbaskalmistute kasutusajaks 150 aastat. Maidla II kivikalme kasutusajaks määratleti 250 aastat ja Maidla I kivikalmel 75 aastat.

158 0 Oodatav eluiga sünnihetkel oli Rõsna külades tagasihoidlik (e 0 = 17,1 aastat), kolme matusepaiga põhjal arvutatud kogukonna suurus oli vastavalt valitud mudelile 16–28 inimest. Suurim, 7–11 inimest, oli kogukond, kes mattis Rõsna- Saare I kääbaskalmistule. Rõsna-Saare II ja Suure-Rõsna kalmistuid kasutanud kogukonnad olid pisut väiksemad, vastavalt 5–8 ja 4–9 inimest. Kasutatud mudelite põhjal võib oletada, et iga kalmistut kasutas reeglina üks perekond, Rõsna-Saare I kalmistut võis kasutada laiendatud perekond või suurem majapidamine. Maidla II noorema rauaaja kivikalmele matnud kogukonna suuruseks on arvutatud 7–9 inimest, mis samuti vastab ühele perekonnale; oodatav eluiga sünnihetkel oli 0 kõrgem (e 0=25,7) kui keskmisel rauaajal Kagu-Eestis. Suhteliselt kõrge oodatav eluiga võib viidata parematele elutingimustele ja madalamale laste suremusele nooremal rauaajal või teatud määral indikeerida laste alaesindatust kalme arheoantropoloogilise materjali hulgas. Põhjused võivad siinkohal olla erinevad: laste matmine mujale, luude purustamine enne matust, tafonoomilised protsessid kalmel ja/või tegevused arheoloogiste kaevamiste käigus, mis muudavad laste ja imikute matused meie jaoks nähtamatuks. Keskmisest rauaajast pärit Maidla I kivikalmet võis kasutada 3–9 liikmeline kogukond ehk samuti pigem üks pere, kes kasutas matusepaika tõenäoliselt vaid 0 lühikese ajaperioodi jooksul. Oodatav eluiga sünnihetkel oli väga madal (e 0=10,9), mis viitab väga kõrgele suremusele ja tõigale, et kogukond ei olnud jätkusuutlik. Kaseküla noorema pronksiaja kivikirstkalmest kogutud arheoantropoloogilise materjali analüüs koos radiosüsinikdateeringutega näitas, kalmet kasutati imikute alternatiivse matusepaigana nooremal rauaajal, seega andis see kalme olulist uut informatsiooni noorema rauaaja matusekombestiku kohta. Radiosüsinikdateeringud koos osteoloogilise analüüsi tulemustega on perspektiivne kombinatsioon uurimaks matusepaiku, kus dateeriv esemeline materjal puudub või on ebamäärane ning kus matuste sooline ja ealine koosseis on ebatavaline. Muinasaegsete kogukondade eluviis, elutsükkel ja matusekombestik on põnevad, kuid keerukad uurimisobjektid, mida me arvatavasti jääme lahti mõtestama veel sajanditeks. Tõenäoliselt võimaldavad regionaalsed arheoloogilised uuringud, mis hõlmavad sama piirkonna erinevaid matusepaiku, meil paremini uurida üleminekut ühelt matusekombestikult teisele ja saada usaldusväärset informatsiooni paleodemograafiliste mudelite koostamiseks.

159 ELULOOKIRJELDUS

Nimi Raili Allmäe Sünniaeg ja -koht 16.12.1963, Tallinn Kodakondsus Eesti

Hariduskäik Alates 2015 Tallinna Ülikool, Humanitaarteaduste instituut, doktoriõpe Alates 2011 Tallinna Ülikool Ajaloo Instituut, doktoriõpe 2005–2007 Tartu Ülikooli filosoofia teaduskond, arheoloogia õppetool, doktoriõpe 1998 Tartu Ülikooli bioloogia ja geograafia teaduskond, üldzooloogia õppetool, MSc (teadusmagister) üldzooloogias ja arengu- bioloogias. 1988 Tartu Ülikool bioloogia-geograafia teaduskond; bioloog (botaanik), bioloogia-keemia õpetaja. Diplomitöö: „Vormsi saare lihhenofloora ja selle analüüs“ 1982 Tallinna 22. Keskkool (Westholmi Gümnaasium)

Teenistuskäik Alates 2017 Tallinna Ülikool, Arheoloogia teaduskogu, teadur 2015–2015 TLÜ Õigusakadeemia, teadusassistent 2014–2014 TLÜ Ajaloo Instituudi keskkonnajaloo keskus (KAJAK), teadusassistent 2007–2013 TLÜ Ajaloo Instituudi arheobioloogia ja muinastehnoloogia osakonna teadur 1998–2007 TLÜ Ajaloo Instituut arheoloogiosakonna teadur 1988–1998 Teaduste Akadeemia Ajaloo Instituudi arheoloogiaosakonna, nooremteadur ja teadur

Uurimisvaldkonnad Füüsiline antropoloogia, arheoantropoloogia, bioarheoloogia

Erialaühendused Alates 2012 TLÜ Keskkonnaajaloo Keskuse (KAJAK) liige Alates 2012 Euroopa Arheoloogide Assotsiatsiooni (EAA) liige

Erialane täiendkoolitus 6.–17.08.2007 Osteoloogiline vanuse määramine ja paleodemograafiline analüüs. Max Planck'i Demograafia Instituudi (Rostock, Saksamaa) ja Lõuna-Taani Ülikooli Kohtumeditsiini Instituudi antropoloogia osakonna koostöös korraldatud kursus Odenses, Taanis

160 CURRICULUM VITAE

Name Raili Allmäe Date and place of birth 16.12.1963, Tallinn Citizenship Estonian

Education Since 2015 Tallinn University, School of Humanities, doctoral studies Since 2011 Tallinn University, Institute of History, doctoral studies 2005–2007 University of Tartu, Department of Philosophy, Institute of Historyand Archaeology, Chair of Archaeology, doctoral studies 1998–1998 University of Tartu, Faculty of Biology and Geography, Department of Zoology and Hydrobiology, MSc, general zoology and developmental biology 1982–1988 University of Tartu, Faculty of Biology and Geography, biologist (botanist), biology-chemistry teacher

Professional experience Since 2017 Tallinn University, Archaeological Research Collection, researcher 2015–2015 Tallinn University, Law School, research assistent 2014–2014 Tallinn University, Center of Environmental History, research assistent 2005–2013 Tallinn University, Institute of History, researcher 1997–2005 Institute of History, State Research Institute, researcher 1988–1997 Institute of History, Estonian Academy of Sciences, researcher

Scientific interests Biological anthropology, archaeoanthropology, bioarchaeology

Professional associations Since 2012 TLU Center of Environmental History, member Since 2012 European Association of Archaeologists (EAA), member

Professional training 06.–17.08.2007 Osteological age estimation and paleodemographic analysis.The Max Planck Institute for Demographic Research (MPIDR), Rostock, Germany in collaboration with Department of Anthropology (ADBOU), Institute of Forensic Medicine,University of Southern Denmark, Odense, Denmark

161 TALLINNA ÜLIKOOL HUMANITAARTEADUSTE DISSERTATSIOONID

TALLINN UNIVERSITY DISSERTATIONS ON HUMANITIES

1. СЕРГЕЙ ДОЦЕНКО. Проблемы поэтики А. М. Ремизова. Автобиографизм как конструктивный принцип творчества. Таллинн: Изд-во ТПУ, 2000. 162 стр. Таллиннский педагогический университет. Диссертации по гуманитарным наукам, 1. ISSN 1406-4391. ISBN 9985-58-135-0. 2. MART KIVIMÄE. Ajaloomõtlemise kolm strateegiat ja nende dialoogisuhted minevikuga (lisades tõlgitud R. Koselleck, J. Rüsen, E. Nolte). Historismi muutumise, arendamise, ületamise probleemid. Tallinn: TPÜ kirjastus, 2000. 201 lk. Tallinna Pedagoogikaülikool. Humanitaarteaduste dissertatsioonid, 2. ISSN 1406–4391. ISBN 9985-58-164-4. 3. НАТАЛЬЯ НЕЧУНАЕВА. Минея как тип славяно–греческого средневекового текста. Таллинн: Изд-во ТПУ, 2000. 177 стр. Таллиннский педагогический университет. Диссертации по гуманитарным наукам, 3. ISSN 1406-4391. ISBN 9985-58-125-3. 4. ОЛЕГ КОСТАНДИ. Раннее творчество В. Каверина как литературный и культурный феномен. Таллин: Изд-во ТПУ, 2001. 142 стр. Таллиннский педагогический университет. Диссертации по гуманитарным наукам, 4. ISSN 1406– 4391. ISBN 9985-58-180-6. 5. LAURI LINDSTRÖM. Album Academicum Universitatis Tartuensis 1918–1944. Rahvus, sugu, sünnikoht ja keskhariduse omandamise koht üliõpilaskonna kujunemist ja kõrghariduse omandamist mõjutavate teguritena. Tallinn: TPÜ kirjastus, 2001. 92 lk. Tallinna Pedagoogikaülikool. Humanitaarteaduste dissertatsioonid, 5. ISSN 1406-4391. ISBN 9985-58-190-3. 6. AУРИКA MEЙMPE. Руccкиe литератoры-эмигрaнmы в Эcmoнии 1918–1940. Нa матepиaлe пеpиoдическoй печaти. Таллин: Изд-во ТПУ, 2001. 165 стр. Таллиннский педагогический университет. Диссертации по гуманитарным наукам, 6. ISSN 1406-4391. ISBN 9985-58-205-5. 7. AIVAR JÜRGENSON. Siberi eestlaste territoriaalsus ja identiteet. Tallinn: TPÜ kirjastus, 2002. 312 lk. Tallinna Pedagoogikaülikool. Humanitaarteaduste dissertatsioonid, 7. ISSN 1406-4391. ISBN 9985-58-239-X. 8. DAVID VSEVIOV. Kirde-Eesti urbaanse anomaalia kujunemine ning struktuur pärast Teist maailmasõda Tallinn: TPÜ kirjastus, 2002. 104 lk. Tallinna Pedagoogikaülikool. Humanitaarteaduste dissertatsioonid, 8. ISSN 1406-4391. ISBN 9985-58-242-X. 9. ROMAN KALLAS. Eesti kirjanduse õpetamise traditsioon XX sajandi vene õppekeelega koolis. Tallinn: TPÜ kirjastus, 2003. 68 lk. Tallinna Pedagoogikaülikool. Humanitaarteaduste dissertatsioonid, 9. ISSN 1406–4391. ISBN 9985-58-256-X. 10. KRISTA KERGE. Keele variatiivsus ja mine-tuletus allkeelte süntaktilise keerukuse tegurina. Tallinn: TPÜ kirjastus, 2003. 246 lk. Tallinna Pedagoogikaülikool. Humanitaarteaduste dissertatsioonid, 10. ISSN 1406-4391. ISBN 9985-58-265-9.

162 11. АННА ГУБЕРГРИЦ. Русская драматургия для детей как элемент субкультуры: 1920–1930-е годы. Таллинн: Изд-во ТПУ, 2004. 168 стр. Таллиннский педагогический университет. Диссертации по гуманитарным наукам, 11. ISSN 1406– 4391. ISBN 9985-58-302-7. 12. VAHUR MÄGI. Inseneriühendused Eesti riigi ülesehituses ja kultuuriprotsessis (1918– 1940). Tallinn: TPÜ kirjastus, 2004. 146 lk. Tallinna Pedagoogikaülikool. Humanitaarteaduste dissertatsioonid, 12. ISSN 1406-4391. ISBN 9985-58-344-2. 13. HEIKKI OLAVI KALLIO. Suomen ja Viron tiedesuhteet erityisesti Viron miehitysaikana vuosina 1940–1991. Tallinn: Tallinnan Pedagogisen Yliopiston kustantamo, 2004. 243 lk. Tallinnan Pedagogisen Yliopiston. Humanististen tieteiden väitöskirjat, 13. ISSN 1406-4391. ISBN 9985-58-350-7. 14. ÜLLE RANNUT. Keelekeskkonna mõju vene õpilaste eesti keele omandamisele ja integratsioonile Eestis. Tallinn: TLÜ kirjastus, 2005. 215 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid, 14. ISSN 1406-4391. ISBN 9985-58-394-9. 15. MERLE JUNG. Sprachspielerische Texte als Impulse für schriftliche Textproduktion im Bereich Deutsch als Fremdsprache. Tallinn: Verlag der Universität Tallinn, 2006. 186 S. Universität Tallinn. Dissertationen in den Geisteswissenschaften, 15. ISSN 1406- 4391. ISBN 9985-58-409-0. 16. ANDRES ADAMSON. Kaitstud ja ilmunud veebiväljaandena. 17. АИДА ХАЧАТУРЯН. Роман В.С.Маканина «Андеграунд, или Герой нашего времени»: Homo urbanis в поле «усреднения». Таллинн: Изд-во ТПУ, 2006. 146 стр. Таллиннский педагогический университет. Диссертации по гуманитарным наукам, 17. ISSN 1736–3624. ISBN-10 9985-58-435-X. ISBN-13 987-9985-58-435-4. 18. JULIA TOFANTŠUK. Construction of Identity in the Fiction of Contemporary British Women Writers (Jeanette Winterson, Meera Syal, and Eva Figes). Tallinn: Tallinn University Press, 2001. 160 p. Tallinn University. Dissertations on Humanities, 18. ISSN 1736-3624. ISBN 978-9985-58-479-8. 19. REILI ARGUS. Eesti keele muutemorfoloogia omandamine. Tallinn: TLÜ kirjastus, 2007. 242 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid, 19. ISSN 1736- 3624. ISBN 978-9985-58-543-6. 20. ÕNNE KEPP. Identiteedi suundumusi Eesti luules. Tallinn: TLÜ kirjastus, 2008. 222 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid, 20. ISSN 1736-3624. ISBN 978-9985-58-574-0. 21. ANNELI KÕVAMEES. Itaalia eesti reisikirjades: Karl Ristikivi „Itaalia Capriccio” ja Amée Beekmani „Plastmassist südamega madonna”. Tallinn: TLÜ kirjastus, 2008. 141 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid, 21. ISSN 1736-3624. ISBN 978-9985-58-543-6. 22. ENE ALAS. The English Language National Examination Validity Defined By its Oral Proficiency Interview Interlocutor Behaviour. Tallinn: Tallinn University, 2010. 232 p. Tallinn University. Dissertations on Humanities, 22. ISSN 1736-3621. ISBN 978-9949- 463-03-9. 23. MERLE TALVIK. Ajakirjagraafika 1930. aastate Eestis: stereotüübid ja ideoloogia. Tallinn: Tallinna Ülikool, 2010. 203 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid, 23. ISSN 1736-3624. ISBN 978-9949-463-31-2.

163 24. TÕNIS LIIBEK. Fotograafiakultuur Eestis 1839-1895. Tallinn. Tallinna Ülikool, 2010. 286 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid, 24. ISSN 1736-24. ISBN 978-9949-463-52-7. 25. HEETE SAHKAI. Teine grammatika. Eesti keele teonimede süntaks konstruktsioonipõhises perspektiivis. Tallinn. Tallinna Ülikool, 2011. 182 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid, 25. ISSN 1736-3624. ISBN 978-9949- 463-98-5. 26. MAARJA VAINO. Irratsionaalsuse poeetika A. H. Tammsaare loomingus. Tallinn. Tallinna Ülikool, 2011. 181 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid, 26. ISSN 1736-3624. ISBN 978-9949-29-017-8. 27. ANNIKA KILGI. Tõlkekeele dünaamika piibli esmaeestinduse käigus: verbi morfosüntaksi areng ja lõplik toimetamisfaas. Tallinn. Tallinna Ülikool, 2012. 222 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid, 27. ISSN 1736-3624. ISBN 978-9949-29-050-5. 28. ELVIRA KÜÜN. Dissertatsioon esitatud kaitsmisele. 29. PEETER KAASIK. Nõukogude Liidu sõjavangipoliitika Teise maailmasõja ajal ja sõjajärgsetel aastatel: sõjavangide kinnpidamissüsteem Eesti näitel ja hinnang sõjavangide kohtlemisele rahvusvahelise õiguse järgi. Tallinn. Tallinna Ülikool, 2012. 631 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid, 29. ISSN 1736-3624. ISBN 978-9949-29-055-0. 30. KADRI SEMM. Milieus in Neighbourhood Place-Making. Tallinn. Tallinna Ülikool, 2012. 210 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid, 30. ISSN 1736- 3624. ISBN 978-9949-29-066-6. 31. AVE MATTHEUS. Eesti laste- ja noortekirjanduse genees: küsimusepüstitusi ja uurimisperspektiive. Tallinn. Tallinna Ülikool, 2012. 260 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid, 31. ISSN 1736-3624. ISBN 978-9949-29-070-3. 32. JELENA KALLAS. Eesti keele sisusõnade süntagmaatilised suhted korpus- ja õppeleksikograafias. Tallinn. Tallinna Ülikool, 2013. 185 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid, 32. ISSN 1736-3624. ISBN 978-9949-29-078-9. 33. KLĀVS SEDLENIEKS. “And Burn Today Whom Yesterday They Fed”: Citizens and State in Montenegro. Tallinn. Tallinn University, 2013. 242 p. Tallinn University. Dissertations on Humanities, 33. ISSN 1736-3624. ISBN 978-9949-29-117-5. 34. МАРИЯ СМОРЖЕВСКИХ-СМИРНОВА. Ингерманландия, Эстляндия и Лифляндия в церковном панегирике петровской эпохи. Таллинн. Таллиннский университет, 2013. 244 стр. Таллиннский университет. Диссертации по гуманитарным наукам, 34. ISSN 1736-3624. ISBN 978-9949-29-119-9. 35. SILLE KAPPER. Muutuv pärimustants: kontseptsioonid ja realisatsioonid Eestis 2008–2013. Tallinn. Tallinna Ülikool, 2013. 241 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid, 35. ISSN 1736-3624. ISBN 978-9949-29-127-4. 36. RAIMONDO MURGIA. The Progressive Aspect in English and Italian: Learning Problems and Remedial Teaching. Tallinn. Tallinn University, 2014. 173 p. Tallinn University. Dissertations on Humanities, 36. ISSN 1736-3624. ISBN 978-9949-29-151-9.

164 37. MARI KENDLA. Eesti kalanimetused: kujunemine, levik ja nimetamise alused. Tallinn. Tallinna Ülikool, 2014. 238 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid, 37. ISSN 1736-3624. ISBN 978-9949-29-157-1. 38. JANIKA KÄRK. Saksa ja eesti keele sagedamate värvingupartiklite võrdlev analüüs. Tallinn. Tallinna Ülikool, 2014. 188 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid, 38. ISSN 1736-3624. ISBN 978-9949-29-160-1. 39. NATALIA ERMAKOV. Эрзянские причитания: традиции бытования и современное состояние. Таллинн. Таллиннский университет, 2014. 175 стр. Таллиннский университет. Диссертации по гуманитарным наукам, 39. ISSN 1736-3624. ISBN 978-9949-29-164-9.

ILMUNUD VEEBIVÄLJAANDENA http://e-ait.tlulib.ee/ 1. ИННА АДАМСОН. Модальный смысл дезидеративности: от семантической зоны к семантической типологии высказываний (на материале русского языка). Таллинн: Изд-во ТЛУ, 2006. 131 стр. Таллиннский педагогический университет. Диссертации по гуманитарным наукам. ISSN 1736-5031. ISBN 978-9985-58-455-2. 2. MARIS SAAGPAKK. Deutschbaltische Autobiographien als Dokumente des zeit- und selbstempfindens: vom ende des 19. Jh. Bis zur umsiedlung 1939. Tallinn: Verlag der Universität Tallinn, 2006. 163 S. Universität Tallinn. Dissertationen in den Geisteswissenschaften. ISSN 1736-5031. ISBN 978-9985-58-469-9. 3. JANIS EŠOTS.Mullā Sadrā’s Teaching on Wujūd: A Synthesis of Mysticism and Philosophy. Tallinn: Tallinn University Press, 2007. 150 p. Tallinn University. Dissertations on Humanities. ISSN 1736-5031. ISBN 978-9985-58-492-7. 4. ГРИГОРИЙ УТГОФ. Проблема синтактического темпа. Таллинн: Изд-во ТЛУ, 2007. 145 стр. Таллиннский педагогический университет. Диссертации по гуманитарным наукам. ISSN 1736-5031. ISBN 978-9985-58-507-8. 5. ДИМИТРИЙ МИРОНОВ. Глагольность в сфере имен: к проблеме семантического описания девербативов (на материале русского языка). Изд-во ТЛУ, 2008. 98 стр. Таллиннский педагогический университет. Диссертации по гуманитарным наукам. ISSN 1736-5031. ISBN 978-9985-58-563-4. 6. INNA PÕLTSAM-JÜRJO. Liivimaa väikelinn varase uusaja lävel. Uurimus Uus- Pärnu ajaloost 16. sajandi esimesel poolel. Tallinn: TLÜ kirjastus, 2008. 257 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid. ISSN 1736-5031. ISBN 978- 9985-58-570-2. 7. TIIT LAUK. Džäss Eestis 1918–1945. Tallinn: TLÜ kirjastus, 2008. 207 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid. ISSN 1736-5031. ISBN 978-9985-58- 594-8. 8. ANDRES ADAMSON. Hertsog Magnus ja tema “Liivimaa kuningriik”. Tallinn: TLÜ kirjastus, 2009. 173 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid. ISSN 1736-5031. ISBN 978-9985-58-615-0.

165 9. ОЛЕСЯ ЛАГАШИНА. Марк Алданов и Лев Толстой: к проблеме рецепции. Таллинн: Изд-во ТЛУ, 2009. 151стр. Таллиннский педагогический университет. Диссертации по гуманитарным наукам. ISSN 1736-5031. ISBN 978-9985-58-654-9. 10. MARGIT LANGEMETS. Nimisõna süstemaatiline polüseemia eesti keeles ja selle esitus eesti keelevaras. Tallinn: TLÜ kirjastus, 2009. 259 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid. ISSN 1736-5031. ISBN 978-9985-58-651-8. 11. LEO LUKS. Ei kogemine nihilismi mõtlemises filosoofia ja kirjanduse ühtesulamisel. Tallinn. Tallinna Ülikool, 2010. 147 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid. ISSN 1736-5031. ISBN 978-9949-463-55-8. 12. JELENA RUDNEVA. „Сказание о черноризском чине“ Кирилла Туровского: опыт лингвотекстологического исследования. Tallinn. Tallinna Ülikool, 2011. 227 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid. ISSN 1736-5031. ISBN 978- 9949-463-92-3. 13. ELO LINDSALU. Naisekuju modelleerimine XX sajandi alguskümnendite eesti kirjanduses. Tallinn. Tallinna Ülikool, 2012. 236 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid. ISSN 1736-5031. ISBN 978-9949-29-024-6. 14. ANTON KÜÜNAL. Специфика оперного либретто как текста: на примере опер на библейские сюжеты (Россия вторая половина XIX b.) Tallinn. Tallinna Ülikool, 2012. 234 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid. ISSN 1736-5031. ISBN 978-9949-29-069-7. 15. EINAR VÄRÄ. Kaubandussidemed Soome suurvürstiriigi ja Eesti alade vahel aastail 1809–1865. Tallinn. Tallinna Ülikool, 2012. 158 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid. ISSN 1736-5031. ISBN 978-9949-29-072-7. 16. INDREK JETS. Lahingu maod. Skandinaavia 9.-11. sajandi kunstistiilid Eesti arheoloogilistel leidudel. Tallinn. Tallinna Ülikool, 2013. 333 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid. ISSN 1736-5031. ISBN 978-9949-29-107-6. 17. MARGUS OTT. Vägi. Individuatsioon, keerustumine ja praktika. Tallinn. Tallinna Ülikool, 2014. 268 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid. ISSN 1736-5031. ISBN 978-9949-29-147-2. 18. MAREK TAMM. Inventing Livonia: Religious and Geographical Representations of the Eastern Baltic Region in Early Thirteenth Century. Tallinn. Tallinna Ülikool, 2009. 201 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid. ISSN 1736-5031. ISBN 978-9985-58-658-7. 19. MELE PESTI. From an Intuitive Metaphor Towards a Working Cultural Model: "Anthropophagy" in Oswald de Andrade's “Anthropophagic Manifesto” and its Development in 20th Century Brazil. Tallinn. Tallinn University, 2014. 208 p. Tallinn University. Dissertations on Humanities. ISSN 1736-5031. ISBN 978-9949-29-163-2. 20. ALARI ALLIK. Multiple Selves of Chōmei / Ren’in. Tallinn. Tallinn University, 2015. 153 p. Tallinn University. Dissertations on Humanities. ISSN 1736-5031. ISBN 978- 9949-29-188-5. 21 LIGITA JUDICKAITĖ-PAŠVENSKIENĖ. Cartoon Subtitling as a Mode of Translation for Children. Tallinn. Tallinn University, 2015. 173 p. Tallinn University. Dissertations on Humanities, 40. ISSN 1736-5031. 978-9949-29-230-1.

166 22 TRIIN VAN DOORSLAER. Conceptualising Translation as an Awareness-Raising Method in Translator Education. Tallinn. Tallinn University, 2015. 85 p. Tallinn University. Dissertations on Humanities. ISSN 1736-5031. ISBN 978-9949-29-236-3. 23. MARIA-MAGDALENA JÜRVETSON. Suomen kielen nauramisverbit koloratiivikonstruktiossa: muoto, merkitys ja tehtävät. Tallinn. Tallinnan Yliopisto, 2015. 194 lk. Tallinnan Yliopisto. Humanisisten tieteiden väitöskirjat, 23. ISSN 1736- 5031. ISBN 978-9949-29-251-6. 24. FRANCISCO MARTINEZ SANCHEZ. Wasted Legacies? Material Culture in Contemporary Estonia. Tallinn. Tallinn University, 2016. 301 p. Tallinn University. Dissertations on Humanities. ISSN 1736-5031. ISBN 978-9949-29-262-2. 25. MARIS SÕRMUS. Nature and Culture in Contemporary British and Estonian Literature: A Material Ecocritical Reading of Monique Roffey and Andrus Kivirähk. Tallinn. Tallinn University, 2016. 160 p. Tallinn University. Dissertations on Humanities. ISSN 1736-5031. ISBN 978-9949-29-264-6. 26. HEILI EINASTO. Eesti balleti rajaja Rahel Olbrei: loometegevus, retseptsioon, pärand. Tallinn. Tallinna Ülikool, 2016. 319 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid. ISSN 1736-5031. ISBN 978-9949-29-265-3. 27. HELEN KÕRGESAAR. Eesti hoidjakeele pragmaatilised erijooned ja dünaamika ning mõju lapse keele arengule. Tallinn. Tallinna Ülikool, 2016. 205 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid. ISSN 1736-5031. ISBN 978-9949-29- 276-9. 28. NATALIA ABROSIMOVA. Развитие эрзянского письменно-литературного языка в контексте культуры. Tallinn. Tallinna Ülikool, 2016. 197 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid. ISSN 1736-5031. ISBN 978-9949-29- 275-2. 29. TATJANA BOEVA. Текст как мыслящая структура: метафизика в.о. пелевина. Tallinn. Tallinna Ülikool, 2016. 192 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid. ISSN 1736-5031. ISBN 978-9949-29-279-0. 30. ANNA RUBTSOVA. Стихосложение Владислава Ходасевича. Tallinn. Tallinna Ülikool, 2016. 132 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid. ISSN 1736-5031. ISBN978-9949-29-280-6. 31. OLGA MJOD. Писатель-протагонист в литературном произведении как частный случай “текста в тексте”. Tallinn. Tallinna Ülikool, 2016. 147 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid. ISSN 1736-5031. ISBN 978- 9949-29-283-7. 32. ANTONIA NAEL. Иллюстрированный «враг»: образ русского «врага» в печатных изданиях межвоенной Эстонии. Tallinn. Tallinna Ülikool, 2017. 204 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid. ISSN 1736-5031. ISBN 978- 9949-29-305-6. 33. MICHAEL AMUNDSEN. Taking it to the streets: getting to the real gritty, gritty through autoethnography. Tallinn. Tallinna Ülikool, 2017. 120 p. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid. ISSN 1736-5031. ISBN 978-9949-29-310-0. 34. MARGE PAAS. Empaatia kunstiteoste kogemisel: fenomenoloogiline analüüs. Tallinn. Tallinna Ülikool, 2017. 133 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid. ISSN 1736-5031. ISBN 978-9949-29-323-0.

167 35. ELLE-MARI TALIVEE. Kuidas kirjutada linna. Eesti proosa linnamaastik aastail 1877–1903. Tallinn. Tallinna Ülikool, 2017. 233 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid. ISSN 1736-5031. ISBN 978-9949-29-318-6. 36. JEKATERINA KOORT. Maastik shanshui 山水 Hiina võimu ja vaimu diskursuses: vaade läbi Song’i dünastia kulgemisõpetuse. Tallinn. Tallinna Ülikool, 2017. 185 lk. Tallinna Ülikool. Humanitaarteaduste dissertatsioonid. ISSN 1736-5031. ISBN 978- 9949-29-327-8.

DISSERTATSIOONINA KAITSTUD MONOGRAAFIAD (ilmunud iseseisva väljaandena) 1. ANNE VALMAS. Eestlaste kirjastustegevus välismaal 1944–2000. I-II. Tallinn: Tallinna Pedagoogikaülikooli kirjastus, 2003. 205, 397 lk. Tallinna Pedagoogikaülikool. ISBN 9985-58-284-5. ISBN 9985-58-285-3. 2. ANNE LANGE. Ants Oras. Monograafia. Tartu: Ilmamaa, 2004. 493 lk. ISBN 9985-77- 163-X. 3. KATRI AASLAV-TEPANDI. Eesti näitlejanna Erna Villmer. Monograafia. Tallinn: Eesti Teatriliit, 2007. 495 lk. ISBN 78-9985-860-41-0. 4. KRISTA ARU. Üks kirg, kolm mõõdet. Peatükke eesti toimetajakesksest ajakirjandusest: K. A. Hermann, J. Tõnisson, K. Toom. Monograafia. Tartu: Eesti Kirjandusmuuseumi Teaduskirjastus, 2008. 479 lk. ISBN 9789949446254.

168